ASR News Issue Volume 15 No. 6 June, 2004

Comments on Vocalabs article: I felt that the Vocalabs recent article "Traps Leading from Good Intentions to Bad Automation" was quite good at presenting many of the central issues that cause self-service to be so much less useful than it could be. The area in which we have disagreement is in our respective views of how the industry should proceed in improving things. The Vocalabs view is that the situation can be substantially improved with better discipline. Giving improved customer service primacy, stressing the importance of the user interface, and, of course, lots of testing, is what Vocalabs advocates. The Vocalabs view is a correct one only if you assume that the basic process that the speech industry is using is a valid one. My view is that the basic design process is flawed. As hard as you may try, you simply can not make chicken salad from chicken poop. The smartest vendor linguists in the world, interviewing end-users with a variety of orientations that have little to do with caller satisfaction, is heading nowhere. Live agents, who have been interacting with thousands of callers know more about the proper user interface than a vendor linguist/designer could ever hope to know. A seasoned agent knows precisely what the correct phrase is and how to say it properly. Yet, much of the industry feels that a linguist/designer that has never spoken to a caller, can somehow figure it out. The existing design process that has been embraced by much of the speech industry vendors is an expensive one that yields marginal results. The live agent is not simply the reviewer. She is the designer. She decides just what is to be said. Also - how it is said (tone, timing inflection, etc.) The agent with the best voice in the call center would be the voice used for recording. Imagine the $ that can be saved here and still get better results. The hours that are spent with "professional voice talent" (that has never talked to a caller), coaching them on how to say a phrase properly, simply dissappears.

Say it like a live agent would: Over-and-over, I run into self-service systems where the phrases that are used are simply not what a live agent would ever say to someone. This is a fundamentally incorrect thing to do. It does not meet the expectations of the caller and puts the caller into a mode in which all of the normal historical behavior of a caller is invalid. The callers response to this is unpredictable. I’ve gotten push-back from the design community that they must provide highly directed phrases in order to get the caller to respond properly. Of course this is the response. Saying a phrase to a caller that that is highly directed, yet does not insult/annoy the caller is something that is difficult to do. It is virtually impossible for a linguist/designer to figure this out. They come up with stuff like: "please tell me the 10 digits of your phone number, saying them clearly one at at time". Listen to what an agent says. It is highly directed, but also subtle and not insulting/irritating to the caller.

Ordering Prescription Drugs: I routinely order prescription drugs via self-service and the supplier is AdvanceRx.com. I happen to use telephone self-service. I use it because I have found that it is the most efficient way for me. I’ve used the web self-service and found it to be less efficient. My experience is that the problems of user interface design are not unique to telephone self-service. Web self-service is generally much worse than telephone self-service. It also is unlikely to ever get much better which means that it will probably never be much of a threat to telephone self-service. The re-order application is a Touch-Tone one. Operation is pretty simple. You enter the prescription number of each drug that is being re-ordered. After all of the prescription numbers are entered, it then obtains the credit card information by either letting the caller use a credit card that was previously used or by entering a new credit card number. It assigns the caller a confirmation number. How would speech improve this? The application does work pretty well already. The opportunity would appear to be in addressing the callers that simply refuse to use telephone self-service. For a speech-enabled self-service application to work for them, it would have to appear as though it were a live agent. The last time that I used the system to order prescription drugs, I noticed that they had a new option that was speech-enabled. Rather than speech-enable the prescription refill application, they decided to speech-enable order status. One of the first things that the speech-enabled order status requested was my "member ID". I did not have a clue what they were asking for. Since I was able to order a refill without proving my "member ID", I found it difficult to understand why I needed to provide it to check the status of an order. I talked to an agent that informed me that the "member ID" was the "ID number" that my insurance carrier had assigned to me. Armed with this knowledge, I tried the order status application again. It was not able to locate me after asking me to say it a couple of times. It then transferred me to a live agent. The agent looked at what had been recognized and confirmed that it had been recognized correctly. What are the people that implemented this order status application thinking about? It certainly doesn’t appear that they have much sensitivity to the needs of the caller.

The Status Quo isn’t going to hack it: Resistance to change is a fundamental characteristic of the world. Newton’s 2nd Law is a basic of law of physics that defines it. We will hang onto what has historically been done. At best, we feel comfortable with incremental tweaking of what we have been doing. For centuries, folks attempted to fly by attempting to emulate birds. They strapped on wings and pursued all sorts of approaches to flapping them in order to fly. It wasn’t until a few brave folks took a totally different approach (basing it on the laws of physics), that we were able to fly. For a number of years, I’ve been advocating a totally different approach to self-service user interface design. What I’ve come to realize is that many people view what I’m saying as being an incremental change from what we have historically been doing. This is not so. An incremental change will accomplish virtually nothing. We presently have a VUI design process in which folks that no nothing about callers, interview customers that no nothing about callers and use the information obtained to meet the needs of these callers that no one knows anything about. Virtually nothing in the training and experience of a linguist would suggest that they are at all appropriate for designing a VUI. Call center managers invariably are totally ignorant of caller’s needs. Have you ever seen a script (provided by the call center management) that is ever really used by the agents? The call center agent should be the designer of the entire VUI. Not just someone that is observed or asked to review the script. VUI design is totally about meeting the needs of the caller. It requires a knowledge of human behavior and experience with real callers.

Is VoiceXML/SALT to blame? Over-&-over, whenever I encounter a speech self-service implementation that is caller satisfaction challenged, I hear the same refrain from the vendors - the customer made us do it. Ignoring the reality of the basic human flaw of instinctly placing the blame with someone else, some truth appears to exist here. An understanding of the needs of the caller rarely appears to be a prerequisite for a customer deciding how speech self-service should be implemented. It is also apparent that few members of the call center management team have much more than a vague idea what the caller is all about. Have you ever seen a live agent actually directly use the script that the call center management provides? The marketing mantra for VoiceXML and SALT is how easy it is to implement speech-enabled applications. This may be a fine message to a point, but it leads to the incorrect conclusion that amateur human factors skills are all that is needed to implement a speech-enabled application. In some ways, this is akin to the medical industry marketing do-it-yourself surgury kits. Tools such as VoiceXML/SALT are fine. Yet they do virtually nothing to address the caller interface and meeting the needs of the caller. Of course I want my surgeon to have good tools to operate on me. More importantly, though, I value his experience and knowledge. By promoting VoiceXML/SALT as we are doing, we are trivializing the most important aspect of speech-enabled self-service - meeting the expectations of the caller.

Just what is the "mainstream": The phrase, "Speech getting into the mainstream" is one that has been frequently used during the last year. The industry press using it quite frequently. When I challenge someone that uses it by asking them to define what they mean, it usually becomes clear that they don’t really have a clue and are just parroting words. When I do get a definition, it often is linked with Microsoft’s entry into the speech area, as though this is going to somehow, magicly transform the industry. The speech industry is about callers. Any sort of measure of "mainstream" would seem to need to relate to the % of callers that routinely like and utilize telephone self-service. By this metric, telephone self-service is not at all in the "mainstream". A significant number of callers intensely dislike self-service and the industry has chosen to, by-and-large ignore this segment of the population, instead focusing on the power users. By doing this, the speech industry has chosen to not get into the mainstream. When we have 90% of the population that likes and uses self-service, then we will be in the mainstream.

ScanSoft says I’m wrong: Shortly after last month’s issue of ASRNews came out, ScanSoft informed me that virtually all of the information that I had included in my articles about them was incorrect. They did not provide any facts that would support their contention, though. I indicated to them that I would be willing to run a retraction if they would provide evidence that disproved the facts that I had previously published. Although they indicated that they would do this, they have not yet done this.

Eckoh Technologies launches speech service for VUE Cinemas 01/06/2004: Eckoh Technologies plc announced that they have delivered a speech-enabled cinema information service to Vue, the UK’s largest multiplex cinema group. The service, which launched in May 2004, provides Vue with a one-number-solution for its group of 40 cinemas across the UK. Vue is the largest operator of modern multiplex cinemas in the country and was formed in May 2003 through the acquisition of Warner Village Cinemas by SBC International Cinemas. The Vue circuit of 42 UK cinemas has 402 screens and over 25 million attendances a year, giving Vue the second largest market share in terms of box office. The cinema information service that has been developed by Eckoh is hosted on their 7,000 line call processing platform in St Albans and Hemel Hempstead, one of the largest speech enabled self-service platforms in the world. Vue customers dial 08712 240 240 to access the service and then using speech recognition they select the cinema they wish to visit either by asking for it specifically or by saying the name of a nearby town or city. They can then say the name of the film they are interested in seeing and the approximate time they would like to visit the cinema and the service will inform them of the most suitable performance times. The caller can then choose to book tickets if they wish. The speech recognition technology means that callers can now access the information they require far quicker than with the traditional touch-tone system, as they no longer have to listen to lengthy lists of options. It also enables mobile callers, an increasingly larger percentage of users, to have a more convenient experience than with touch-tone as they no longer have to keep moving the handset away from their face to make a selection. Contact Eckoh Technologies plc Martin Turner, Chief Executive Officer Nik Philpot, Chief Operating Officer Tel: 08701 100 700 Roland Jones, IT Director at Vue

CyraCom International Partners with Envox Worldwide and Nuance Communications to Deliver Enhanced Language Services to the Healthcare Market: CyraCom International, Inc announced that it is partnering with Envox Worldwide to develop and deploy CyraCom's phone-based interpretation service. This solution, known as Transparent Language Services, enables better communication between healthcare providers and patients with limited proficiency in English. The new solution incorporates the Envox Communications Development Platform and Nuance 8.5, Nuance's (Nasdaq: NUAN) speech recognition engine. This new product enhancement adds to CyraCom's complete range of healthcare communication services that includes on-site and phone-based interpretation services in more than 150 languages. As the number of ethnically and culturally diverse patients in the US continues to rise each year, errors in non-professional medical interpretation are common and can have serious clinical implications. To ensure fast, accurate diagnoses, healthcare providers need competent, full service medical interpretation and translation services. To address this demand, CyraCom pioneered the creation of transparent multi-lingual interpretation and translation services. CyraCom offers its customers the benefits of efficient and rapid service through its partnerships with leading technology providers like Envox Worldwide and Nuance. CyraCom's newest service enhancement features advanced speech and telephony capabilities, which translate to speed and cost benefits for its customers. First, through Nuance 8.5, CyraCom offers a speech-enabled language selection menu allowing healthcare providers to simply speak the name of the language they require. This saves time as healthcare providers do not have to manually look up the required language code and enter it on the keypad. Additionally, by utilizing the Envox platform's advanced call control functions, CyraCom developed a new innovative queuing feature to provide interpreters more rapidly, in many cases before the client is ready to engage the call. This not only means faster service for the customer but also reduced costs. CyraCom's Transparent Language Services give callers the benefit of faster and more efficient phone-based interactions, while helping healthcare provider companies improve customer satisfaction and lower costs. Powered by Nuance 8.5, Nuance's speech recognition software, Cyracom's offering delivers highly accurate service and is scalable for large call volumes. Additionally, by utilizing Envox's graphical programming environment, Cyracom gains a streamlined development methodology for quickly designing new product features and expanding customer benefits. CyraCom International provides Transparent Language Services on demand, 24 hours a day, 365 days a year in over 150 different languages accessible through an 800 number. This service can be programmed seamlessly into an organization's current PBX and telephone network for quick one button access. This eliminates the need for operator assisted calling, though this type of connection can be requested. The service can also be accessed using a variety of telephonic devices, including handsets, speakerphones or even cellular and wireless telephonic equipment. For additional convenience, CyraCom offers a dual handset phone, known as the CyraPhone. This unique phone is pre-programmed for instant access. Equipped with two handsets, users are able to communicate face-to-face while protecting the confidential integrity of the conversation. CyraCom provides a complete spectrum of communication services in more than 150 languages. Interpretation is available on-site and telephonically (over standard phone lines or using the company's patented telephone technology) as well as translation, staff training, qualification, and assessment services. CyraCom's communication specialists are trained in medical-specific terminology and concepts, enabling CyraCom to better meet the critical needs of healthcare professionals. 800-713- 4950 www.cyracom.com. The Envox Communications Development Platform is an open, standards-based platform for creating, deploying and managing the voice solutions that enterprise customers and service providers require today. Key features include a graphical programming environment, an embedded VoiceXML browser, numerous integrations for leading telephony and speech products, and powerful management tools for easily configuring, monitoring and managing large-scale deployments. It is a true multi-modal communications development platform that greatly reduces the complexity inherent in creating and managing Voice Solutions and integrating them into diverse telephony, data, web services and back office environments. www.envox.com. www.nuance.com or call 1-888-NUANCE-8. Envox Media Information - Americas, contact Gretchen Clarke, +1 508 871 7603, gclarke@envox.com or Sarah Williams, +1 617 585 2780, swilliams@brodeur.com Envox Media Information - EMEA, contact Franca Zucca, +46 8 562 560 65, franca.zucca@envox.com Envox Media Information - Asia/Pacific, contact Thanaletchumy Arujanan, +603 62035882, thana@envox.com CyraCom Media Information - America, contact Neva Rountree or Kay Cavender, Rountree Group, 770-645-4545 Nuance Media Information - America, contact Nancy MacGregor Hill, Eastwick Communications, 650.596.6501, nhill@eastwick.com Michael D. Greenbaum, CEO of CyraCom.

CSX Transportation Keeps Trains Moving With Speech-Enabled Crew Management Application From Intervoice: Intervoice, Inc. (Nasdaq: INTV - News), announced that CSX Corporation (NYSE: CSX - News), the largest rail network operator based in the East, is now offering a speech-enabled crew management application powered by the Omvia platform that provides CSX Transportation (CSXT) train conductors with job assignments while complying with union agreements and optimizing dispatch resources. From unit trains of coal to truckloads of parcels on flatcars, CSXT is utilizing speech recognition technology to chart a path in becoming the most progressive freight transportation company in North America. The project was completed and fully deployed in the Company's previous fiscal year which ended February 29, 2004. Spanning some 23,000 miles in 23 states, the District of Columbia and two Canadian provinces, CSX Transportation serves more North American ports than any other U.S. rail network operator. An important facet of daily operations lies with CSXT dispatchers who focus on logistics for train and engine employees to ensure that trains keep moving. Because train and engine employees are union, select assignments are available based upon their seniority. Dependent on which union employees are members, 20 different union agreements provide rules that need to be followed regarding job assignments. Prior to the deployment of this application, all CSXT train and engine employees had to speak with a crew dispatcher to make a change from one assignment or job to another. The recently-deployed, speech-enabled application from Intervoice incorporates all of the agreements and their respective rules, resulting in a complex-made-easy, self-service application that currently handles up to 1,000 seniority moves a day. Train and engine employees call into the speech-enabled system and dictate to the system what kind of job they are looking for and the system tells them what jobs are available; then they select the job -- all via speech recognition. Employees are managing assignment and job moves by calling into the Intervoice system. This eliminates the need for human involvement; requests are handled instantaneously and consistently, allowing dispatchers to manage more complex functions. Before CSX had this application, when an employee would call in and work through the crew dispatcher to make a seniority move, it was taking an average of 10 minutes. Currently, the process is taking an average of two minutes and 35 seconds per move, reducing average call durations by 74 percent. CSX Corporation, based in Jacksonville, Fla., owns the largest rail network in the eastern United States. CSX Transportation Inc. and its 34,000 employees provide rail transportation services over a 23,000 route-mile network in 23 states, the District of Columbia and two Canadian provinces. CSX Corporation also provides intermodal and global container terminal operations through other subsidiaries. http://www.intervoice.com . Tim Male, Assistant Vice President of Crew Management

National Rail Enquiries extends electronic access: National Rail Enquiries plans to introduce new ways of accessing its services to attract more users and cut operational costs. The organisation will offer text and automated speech recognition services by the end of the year, to add to the existing online journey planner, live departure boards relaying information from stations to its web site, and a WAP service. The automated speech system will use a different phone number, but at busy times a message on the main enquiry number will alert users to the alternative. Technical details have yet to be fully determined, but contracts will be let in the next few weeks. The text service, developed by Kazoom, will allow passengers to register on the web site and receive train details by text, at a cost of 25p per message. By the end of the year, a fully automated service will allow passengers to text enquiries about trains and receive a text response. 'If there's something serious on the network like snow or flooding, call levels can double or even quadruple. That's when the text service will be optimised,' National Rail Enquiries receives 95 million contacts a year, with 50 million handled on the phone. Some 10 per cent of calls are now answered in India under an outsourcing deal with BT and Ventura. chief executive Chris Scoggins. National Rail Enquiries

Virgin Atlantic Flying High with New Voice-Based Flight Information Service from Nuance: Virgin Atlantic's (London, UK) customer service will take off to new heights, thanks to a contract awarded to Nuance (Nasdaq:NUAN). After a competitive bidding process, Nuance was chosen to implement a speech-based automatic flight information line in both the United States and United Kingdom. The dedicated voice-based flight service will respond to customers' inquiries and enable Virgin Atlantic fliers to rapidly obtain information about flight departure and arrival times. The speech-based system is more flexible and intuitive than Virgin Atlantic's previous touch-tone phone system and will handle more than half a million customer calls a year. The new system will also allow Virgin Atlantic contact center agents to spend more time with customers that have more complex inquiries. The information line will be customized for Virgin Atlantic to ensure the voice on the other end is personal to the Virgin Atlantic brand by using Nuance's Persona Design Service. Virgin Atlantic is using the Nuance Voice Platform. www.nuance.com or call 1-888-NUANCE-8. Kim Pharro, telecommunications team leader at Virgin Atlantic

Montreal's Public Transportation Organization Selects Nuance Vocalizer 3.0 Canadian French Text-to-Speech Software: Nuance (Nasdaq:NUAN) announced that La Societe de Transport de Montreal (STM), Montreal's public transportation organization, has deployed Nuance Vocalizer 3.0 text-to-speech (TTS) software in its bilingual door-to-door transportation information line. This application enables disabled passengers to call into STM's automated traveler information line and hear needed reservation and route data spoken back to them in a natural, conversational manner via Nuance Vocalizer text-to-speech in Canadian French or English. STM has also deployed two new and very successful features that use Nuance Vocalizer: AUTOBUS, which allows users to obtain the next three arrival times at any given bus stop, and STM.INFO, which provides a variety of information to customers. STM is now planning to offer riders more voice-automated services, including detailed itineraries and route change information. Nuance Vocalizer converts text in a database -- such as an airline's flight schedule information or directory assistance listings -- into spoken words, allowing companies such as STM to automate the delivery of dynamic information over the phone. Moreover, Vocalizer, by reducing the number of recorded prompts necessary in an application, can lower the costs associated with recording the prompts needed for automated speech applications. Professional prompt recordings can cost more than $800 per hour, with an application taking weeks to record and edit in a professional studio. Therefore, a reduction in the number of recorded prompts can have a significant impact on application cost and deployment time. Nuance Vocalizer is deployed at STM with the Cisco IP Interactive Voice Response (IP IVR) product, an integral component of Cisco IP Communications, a comprehensive system of powerful, enterprise-class solutions -- including IP telephony, unified communications, IP video and audio conferencing, and contact center -- that is enabled by Cisco AVVID (Architecture for Voice, Video and Integrated Data). Cisco IP Communications increases STM's competitive advantage by delivering a measurable return on investment (ROI) through improving efficiency, boosting productivity, and enhancing customer loyalty. Nuance now offers a total of 18 TTS languages, including Nuance Vocalizer 3.0 North American, UK and Australian English, Canadian French and Latin American Spanish, and -- via Nuance partners -- Brazilian Portuguese, European French, European Spanish, Dutch, Danish, German, Italian, Korean, Japanese, Cantonese Chinese, Mandarin Chinese, Norwegian and Swedish. STM joins more than 200 companies using Nuance Vocalizer, including Countrywide Financial and Washington State Department of Transportation. Partners benefiting from Nuance Vocalizer include Amcom Software, Edify, Gold Systems, Intervoice, ISx, Nortel Networks, Workforce Technologies and Versay. www.nuance.com 1-888-NUANCE-8. CONTACT: Eastwick Communications, Becky Quinlan, 650-480-4032 becky.quinlan@eastwick.com Serge Belanger, information systems manager for STM

Syntellect's Automated Solution at Camelot: Syntellect, a division of Enghouse Systems Limited (TSX: ESL) announced that Camelot (Berkshire, UK) the national lottery operator, has implemented a Vista-based solution to provide enhanced automated services for its customer National Lottery Line. Each day, The National Lottery receives thousands of winning number enquiry calls from players. As part of a continued drive to improve customer service and response times, Camelot has introduced a new automated service to extend access to the winning numbers on all its draw-based games - Lotto, Lotto Extra, Lotto HotPicks, Thunderball, Daily Play and EuroMillions. This includes both historical as well as recent winning numbers and prize amounts. This service is of critical importance to ensure that during peak call demand following a draw, players can access the National Lottery Line. With up to 95 percent of callers requesting winning numbers information only, the automated service will provide timely and accurate information on both recent and historic draws, thus freeing representatives to spend more time on other calls. Camelot sees the provision of an automated information service for players as one of the many initiatives that will provide value added solutions to the National Lottery Line service. Camelot Group plc is the operator of The National Lottery and is committed to raising money for the Good Causes designated by Parliament. Camelot is not responsible for distributing or awarding these funds. To date, £15.8 billion has been raised for Good Causes by The National Lottery, benefiting 171,169 individual projects UK-wide in the biggest programme of civic and social regeneration since the 19th Century. The National Lottery has given away over £22.5 billion in prizes and created more than 1,650 millionaires or multi-millionaires since 1994. Camelot runs one of the most cost efficient lotteries in the world, using only 5 per cent of total revenues in operating costs. At 40 per cent of total sales (28 per cent to Good Causes and 12 per cent in lottery duty to the Government), Camelot returns a higher proportion of lottery revenue back to society, than any other lottery in the world. Camelot is committed to the highest standards in player protection and social responsibility in both the retail and internet environments, and was ranked in Business in the Community's 2003 Top 100 Companies That Count, in terms of social responsibility. www.camelotgroup.co.uk and www.national-lottery.co.uk. 1.800.788.9733 or in the UK on +44 (0) 1628 897 500 and say "Syntellect www.syntellect.com. Contacts: Ben Rosier/ Miranda Pugh, Media Relations Managers Camelot Press Office: 020 7632 7511 Philip Hamilton Marketing Manager
Syntellect Ltd International Office E-mail: phamilton@syntellect.co Tel: +44 (0) 1628 897 531 Tom Scott, Head of Communication Services at Camelot

TeleTimecard Announces New Time and Attendance Solution: TeleTimecard.com LLC (Rochester, MN) announced the launch of a new service that allows companies to track and record the attendance of their mobile workforce via telephone. Using a standard touch-tone phone, remote and dispersed workers can check-in and check-out from a client's location or place of residence and report special hours, activity codes, mileage or expenses. Companies also have the capability to broadcast special messages to employees when they call in. Setup and administration of TeleTimecard can be accomplished in as little as 15 minutes. After registering, companies simply create new employee and client lists. Then, using the TeleTimecard Setup Wizard, companies determine how complex their dynamic phone scripts will be. Employees can begin using TeleTimecard immediately. All controls and real-time reporting are accessible through an intuitive web-based interface. With TeleTimecard, companies can monitor and record employee attendance in real-time via the Web. They can provide documented evidence of employee time and attendance via Caller ID tracking. More importantly, companies can eliminate paper timecards and the staff time required for manual data entry. TeleTimecard has signed up clients in the non-medical, senior care and home health industries and is targeting their solution to a variety of other industries including field service/repair, janitorial/cleaning services, security services, transporation companies as well as the government/public sector. TeleTimecard is offering the solution at a fixed monthly fee plus per minute charges. www.teletimecard.com Contact: John Pacchetti TeleTimecard.com Phone: 507-285-1700 E-mail: john@teletimecard.com

Roxio Selects Tuvox To Deliver World-Class Call Center Automation: TuVox announced that Roxio, Inc. (NASDAQ: ROXI) - provider of the best selling digital media software and owner of Napster, the world's most recognized brand in online music - has selected TuVox to provide world-class call center automation that better services the needs of its growing customer base. With TuVox's speech application technology, Roxio aims to automate up to 30 percent of its calls, keeping customers well attended to, while reducing costs associated with answering routine customer service inquiries and more complex support calls. Roxio expects a significant cost savings by investing in TuVox's technology. In addition to reducing costs and supporting growth, TuVox's call center automation solution also provides Roxio with an advantage over most of its competition. Unlike many of its competitors who limit access to customer service conducted via phone, Roxio is able to offer its customers a streamlined, hassle-free call-in experience at no charge. TuVox's mission is to enable call centers to offer voice self service that is so effective that callers will freely choose automation over a live agent. Built-in Perfect Agent capabilities automatically incorporate best practices into every dialog - making high-quality speech applications accessible and affordable. TuVox customers include companies such as Activision, TiVo, MCI and Definity Health. www.tuvox.com. Contact: Jennifer Balsam, SHIFT Communications, 415-591-8405 jbalsam@shiftcomm.com Gary Schultz, director of international support, Roxio

Pegasus Solutions Deploys Voxify Automated Agents to Enhance Services for Utell by Pegasus Member Hotels

Voxify(Alameda, CA) announced that Pegasus Solutions, a leading worldwide provider of reservations-related services and technology to the hotel industry, has successfully deployed Voxify Automated Agents to enhance the telephone services it provides to more than 4,200 Utell by Pegasus member hotels. Pegasus had investigated speech-driven call center automation previously, but the cost and risk of developing custom applications and the lengthy deployment cycles convinced Pegasus that such an approach wouldn't be cost-effective for the firm or its clients. That changed after Pegasus met Voxify. Following an internal trial, Pegasus rolled out Automated Agents for Utell member hotels. Today, powered by Voxify’s patented Conversation Engine software platform, Automated Agents are helping Pegasus serve confirmation and cancellation calls, with a view to extending this service to handle end-to-end reservations. Voxify Automated Agents use the Conversation Engine to model the intelligence and personality of a human agent, allowing them to engage callers in sophisticated dialogue and perform advanced customer service functions, including reservations, sales, and account requests. Pegasus initially allowed Voxify to answer only 50 percent of their confirmation and cancellation calls. Voxify deployed within eight weeks and achieved a call completion rate of over 95 percent. With Voxify handling the majority of confirmation and cancellation calls, Utell member hotels are benefiting from enhanced automated services. Voxify was founded by MIT technologists and enterprise software veterans. Voxify has a growing list of skilled partners around the globe. Contact: Hollis Chin, Voxify 510.545.5015 hchin@voxify.com Tim Beyers, Impact! Studio 720.981.8818 tbeyers@getimpact.com Nasir (Kass) Kassum, Senior Vice President of Data and Voice Services at Pegasus Solutions.

Intervoice Announces Suite of 10 Process-Based Applications for VoiceXML and SALT Environments: As part of its strategy to provide voice solutions that assist enterprises and carriers to exceed end- user expectations through the use of speech technologies, Intervoice, Inc. (Nasdaq: INTV), announced the general availability of its Omvia Voice Express applications suite which supports both SALT and VoiceXML. Transforming critical business processes into conversations, the 10 available and flexible Omvia Voice Express applications enable small to large enterprises and carriers to connect people and information while reducing upfront development costs, deployment timeframes and the complexity associated with application development. Developed with customer service delivery processes that address business problems, Intervoice identified four horizontal process areas where voice technology is best suited: general information management, customer acquisition, billing and customer care. The General Information Module allows customers to browse through products and services of interest while enabling them to generate information requests for future purchase considerations. The Customer Acquisition Module enables customers to order the product or service of their choice while also allowing their order to be fulfilled and confirmed. With the Billing Module, end users are provided with a method to pay for goods and services through the generation, collection and confirmation of relevant billing information. Finally, in the event a customer needs service, the Customer Care Module provides service requests, service scheduling, service confirmation, returns/exchanges and account information requests. Depending on the need for the application, platform and target audience, Omvia Voice Express applications from Intervoice can be used as application templates, turn-key solutions, applications as a service (ASP), configurable applications and developer applications. By having this flexibility, companies can harness the power of early adoption by the developer community while taking advantage of voice applications that suit the needs of both the large enterprises and service providers as well as those of small-to-medium sized businesses. The comprehensive Omvia Voice Express application portfolio includes: General Information Management Module -- Omvia Communicator -- Offers callers fast and efficient access to other people, places and information resources from any telephone device at anytime, using simple natural voice commands that combine the most advanced Voice Activated Dialing functionality available in the industry. Communicator is built on a proven architecture that fully supports advanced speech recognition, provides standard interfaces to the telephony network, and provides a speech-activated interface to name dialing and corporate and shared address books. -- Omvia Locator -- Provides advanced, speech-enabled locational services by city, state, or zip code. Additionally, Omvia Locator with Cross Street Finder makes finding a retail outlet, kiosk, or other services easier through speech-enabled cross-street identification. The Locator module provides generic, easy-to-configure location-based functionality that can be customized to meet a wide range of organizational requirements. The solution can be deployed, for example, to help customers find a store or a branch office from a caller-specified location such as cross streets. Locator makes giving consumers directions to any business simpler and less costly. -- Omvia Rate Quote -- Offers "fast quotes" to give callers instant access to popular interest rates. The application captures information such as property value, loan amount, property type, loan type and term to provide custom quotes to callers and provides a library of mortgage questions that can be easily selected by the financial institution. Omvia Rate Quote also allows callers to perform complex calculations with quick navigation. -- Omvia Name & Address Capture -- Accurately captures first, middle and last names, zip code, city, state, street number, street name and apartment number. Name & Address Capture shortens call durations using reverse phone number lookup functionality based on a 10-digit phone number of Automatic Number Identification (ANI) and provides integrated third-party address verification and notifies callers of invalid primary and secondary address information. -- Omvia Notification -- Available as a service, Omvia Notification, provides reliable and cost-efficient, voice-enabled emergency event notifications that combine sophisticated monitoring and detection with advanced push technology to virtually any wireline, wireless, Internet-enabled or communication device, including voice telephones, pager, email, PDA, SMS-text messages or a combination of devices. Customer Acquisition Module -- Omvia Funds Transfer -- Advises callers about account eligibility, balance minimums, transaction limits, and insufficient funds. The application verifies transfer information before the transaction is completed, allows callers to change particular pieces of transfer information without restarting the transaction and supports transfers from external financial institutions. Funds Transfer also automatically provides account balances after transfer is completed. Billing Module -- Omvia Account Payment -- For credit card holders, callers can pay the current balance, last statement balance, minimum payment or custom amount. For mortgages and loans, callers can choose the monthly payment or a greater custom account. The application supports mortgage and loan payoffs, payments from external financial institutions and the setup of recurring automatic payments and scheduled payments to ease the payment process. Customer Care Module -- Omvia Survey -- Allows businesses to gauge customer satisfaction and conduct televoting initiatives through voice automation. Once a survey or televoting process is created, Survey answers inbound calls and queries callers for responses. Companies have the ability to capture recorded customer comments for further analysis. The user interface uses text-to-speech or pre-recorded prompts and records callers' verbal comments. Multiple surveys can run simultaneously and the caller will be directed to specific ones based on DNIS or by a Survey Identifier related to their specific caller account. -- Omvia Identifier -- Delivers an efficient, cost-effective way to meet the needs of clients who lose, forget or need to change their password or Personal Identification Number (PIN). The PIN and Password Reset application from Intervoice can be deployed to allow callers and subscribers to quickly and securely manage these key self-service system identifiers. Identifier provides a comprehensive reset solution, including a consistent Voice User Interface (VUI), caller authentication, and back-end Reset functionalities that are easily customized for use in a wide range of industries. The Identifier solution allows callers to log in and be identified and authenticated, to change their PIN or password, and then to enforce that change within the automated system, giving companies powerful flexibility in managing customer ID changes. -- Omvia Account Information -- Allows callers to retrieve transaction history, balances and status of a specific check while providing filters for reviewing transaction history based on date ranges and transaction types. The application also includes optional warnings for account balance minimums and activity thresholds. Intervoice has made speech easy through the definition of industry- specific grammars and reusable application objects that are used as the building blocks of Omvia Voice Express. Through this approach, Intervoice is able to rapidly provide the many industry segments with specific voice solutions that meet end-users' needs. The Intervoice implementation approach is a four-step process that enables enterprises to select an Omvia Voice Express application of choice, configure it to their needs via tools supporting both .NET and J2EE environments, and choose voice response technologies, through support of the Omvia Voice Framework that supports both .NET and J2EE environments and technologies such as: -- Microsoft Speech Server 2004, including support for the Intervoice Telephony Interface Manager (TIM) and Intel TIM, -- Intervoice Media Server for Voice XML and -- Intervoice Computer Telephony Integration (CTI) options for call center deployments. Additionally, Omvia Voice Express was designed to ensure that enterprises are not "locked in" to technology so that companies can choose an infrastructure, such as those provided by industry leaders like HP and Dell. Powered by Microsoft SQL Server, Omvia Voice Express applications can provide enterprises with valuable operational and business insight. Detailed call logs and out-of-box reports help managers analyze system usage, efficiency and performance. Applications can be accessed by calling 1-866 TRY-INTV. Intervoice solutions employ data object modeling and an advanced application control environment, which simplify the integration of back-end systems and protect existing infrastructure investments. These applications also feature a modular construction that allows developers to add or remove functionality without reworking the entire system. Intervoice speech applications incorporate the reporting and business intelligence features needed to measure system performance, caller usage and behavior. Intervoice offers the Agility Quick Start package of services to support the customer's implementation of the Omvia Voice Express applications. The Quick Start package leverages the Professional Services Agility Suite, a portfolio of services designed to address fundamental business processes as well as how to build, implement and optimize speech applications to best suit the customer's needs and leverage their return on investment (ROI). Intervoice has sold more than 23,000 systems worldwide including Ameritrade, Amtrak, Citibank, MasterCard, O2, Rogers Wireless, SBC, Travelocity, Verizon and Vodafone. A Microsoft Certified Partner, Intervoice is headquartered in Dallas with offices in Europe, the Middle East, South America, Africa and Asia-Pacific. www.intervoice.com Contact: stephanie.leonard@intervoice.com 972-454-8231

Vocantas Announces Compliance OnCall to Address Privacy Legislation: Vocantas Inc. (Ottawa, ON) announced the release of Compliance OnCall solution. Compliance OnCall, is designed to ease the process of validating customer consent required to meet tough new privacy laws in Canada and the U.S., including the Canadian Personal Information Protection and Electronic Documents Act (PIPEDA). PIPEDA and other similar privacy legislation requires that personal information held by an organization that may identify an individual be accurate and up-to-date. In addition, the purpose for which such information is used, collected and disseminated must be disclosed and agreed to by the individual from whom it is collected. These requirements now apply to every public and private organization that handles personal information in the course of commercial activities – from the corner video rental store to the largest multi-national corporation. Vocantas’ Compliance OnCall helps enterprises accomplish these requirements by automating the process for bringing consent information up to current standards. The Compliance OnCall application also can be used for other customer relationship management purposes, thereby increasing its value beyond a one-time expenditure. The cost savings of automating the consent process can be substantial. For example, a hydro utility that is now using Vocantas had previously mailed out PIPEDA consent forms to one third of their customer base, representing approximately 30,000 customers. Part way into the process, the utility determined that without taking into account postage and staff time required to process the returned responses from only a part of its customer base, the production costs of the forms alone would have paid for the entire Vocantas solution. More importantly, Compliance OnCall would have allowed the entire process to happen seamlessly from beginning to end, to encompass the entire customer base rather than just a part of it, to provide compliance reporting tools for the chief privacy officer, and to provide a re-usable automated application allowing the utility to easily contact its customers for other purposes. Vocantas Inc. is the solution provider of the OnCall customer service suite of solutions with experience in the application of computer telephony and speech recognition in business environments and a provider of turnkey touch-tone (DTMF) and voice-enabled customer service solutions. The company’s product offering includes interactive voice response (IVR) systems, proven computer telephony integration (CTI) applications and Text-to-Speech and speech recognition environments. www.vocantas.com Contact: Randy Brown, Director of Sales & Marketing, Vocantas Inc. Tel: (613) 271-8853 ext. 596 E-mail: randy.brown@vocantas.com

SupportSoft and Tell-Eureka Team to Make Telephone Self-Service More Efficient: SupportSoft, Inc. (Nasdaq: SPRT - News), and Tell-Eureka Corp. announced the planned integration of SupportSoft's VoiceAssist product with Tell-Eureka's LevelOne Virtual CSR solution. The combined solution is expected to extend to broadband service providers the ability to offer their customers telephone-based self- service that more precisely and quickly resolves technical problems on the first call, leading to faster and more personalized customer service. Integration of SupportSoft's patented SmartIssue diagnostics and knowledge automation functionality with LevelOne Virtual CSR allows cable and broadband subscribers to solve problems themselves with the help of Tell-Eureka's virtual agent. LevelOne uses speech recognition technology to intelligently and automatically understand the caller's computing environment and troubleshoot complex technical problems, enabling the service provider to more efficiently utilize valuable support staff. Additionally, by combining knowledge automation capability through use of SupportSoft's Knowledge Center Suite with Tell-Eureka's broadband voice knowledge bases, customers can even receive automated, "one click fixes" directly at their computer desktops to solve the problems. For broadband subscribers who prefer to use the telephone for customer service, Tell- Eureka's LevelOne Virtual CSR can already be well on the way to solving the problem once the caller inputs a sixteen digit code generated by SupportSoft's VoiceAssist software. LevelOne Virtual CSR, a voice self-service solution for contact centers, automates first-level customer support using conversational speech recognition. LevelOne reduces per-call costs and improves overall customer satisfaction. VoiceAssist utilizes patented SmartIssue technology to automatically summarize the user's key system settings and translates these into a numbered sequence, even when they are offline, so the data transaction between the caller and the customer service center can be easily and accurately conveyed. The caller can either enter the SmartIssue code sequence into a voice self- service solution, like Tell-Eureka's, or give it verbally to a support professional. In both instances, VoiceAssist reduces the frustration from the often lengthy and traditional phone-based discovery process, resulting in more satisfied customers. SupportSoft is a leading provider of Real-Time Service Management (RTSM) software designed to accelerate and automate enterprise technical support, customer service and IT infrastructure management. Enterprises that have purchased our products and services include: ADP, Bank of America, Cisco Systems, IBM, Procter & Gamble, Siebel Systems, Sony and Thomson Financial. Managed service providers that have purchased our products to provide outsourced services to enterprises include: Accenture, ACS, CompuCom, CSC and IBM Global Services. Digital service providers incorporating our software into their service offerings include: Adelphia Communications, BellSouth, Charter Communications, Comcast Communications, Cox Communications, TeliaSonera and UPC. www.supportsoft.com. Tell-Eureka is a developer of multi-channel, self-service customer support applications that combine speech recognition and Web technologies. 303-545-2280 www.telleureka.com. Scott Dale, SupportSoft's vice president of engineering, Zor Gorelov, Tell-Eureka's CEO

TuVox Enhances Its Automated Call Center Solution With New Capabilities: TuVox (Los Altos, CA) announced the availability of TuVox CVR 4.0, a major new product release designed to dramatically improve the way speech applications are developed, deployed and maintained. By using TuVox CVR 4.0, which expands TuVox's built-in Perfect Agent services and SmartGen tools, companies can quickly and easily deploy highly customized automated speech applications. TuVox CVR 4.0 builds on the success of previous TuVox offerings by expanding the types of calls that can be automated, allowing call centers to automate calls of any type or complexity, including knowledge calls, routing calls and now sophisticated transactional calls. Additionally, TuVox CVR 4.0 enhances the built-in Perfect Agent feature, which benefits callers by creating an automated experience that is more efficient and easier to use. Incorporating a new suite of SmartGen tools, TuVox CVR 4.0 enables IT organizations to create, maintain and enhance their own speech applications, reducing dependency on expensive consultants. *More automation: New DataBrowser feature expands call types that can be automated to include sophisticated transactional calls, by providing callers with voice access to complex tabular data across a variety of databases. *More conversational: Enhanced Perfect Agent capabilities include a new Predictive Recognition feature that allows customers to fine-tune their speech applications before they are deployed. By predicting and improving handling of "mis-recognized" words or phrases, this revolutionary feature eliminates the need to use callers as virtual "crash test dummies" - ensuring a better initial caller experience. *More control: Enhanced suite of SmartGen tools designed to automatically generate and update speech applications. A new Real-time Application Monitoring feature alerts call-center staff to problems and updates in real time. TuVox customers include companies such as Activision, TiVo, MCI and Definity Health. www.tuvox.com. Contact: Jennifer Balsam, SHIFT Communications 415-591-8405 jbalsam@shiftcomm.com

Traps Leading from Good Intentions to Bad Automation: Nobody ever makes a goal of providing bad customer service. Yet, judging from consumer complaints and the generally poor quality of automated customer service systems, bad service happens far more often than anyone in the contact center industry wants to admit. Building a truly outstanding automated customer service system takes hard work and discipline. Creating a bad system merely takes hard work. There are a number of pitfalls which can derail good intentions and lead to a poor end result. Automating Only to Save Money: Customer service automation is an important technique for controlling costs, but just as importantly, it can lead to better customer service. We've found that for a well-designed system, customers actually would rather use the self-service option than speak to an agent. This leads to a triple win: the customer gets better service, the company saves money, and the project manager gets promoted. But when a company approaches building customer service automation solely from the perspective of saving money, the opportunity to improve service is often missed. The end result is often more expensive, since customers are less willing to use an automated system which doesn't serve their needs. Measuring Success with the Wrong Criteria: Determining the success or failure of a project is crucial, and the measurable goals of the project will steer nearly every design and implementation decision. Often, however, the success criteria do not really reflect the business goals of the project, or reflect them in a distorted way. For example, if the business case was built around both saving money and improving service, but only metrics related to saving money are measured, then the goal of improving customer service might as well not exist. Not Measuring Customer Impact: Even if improved customer service isn't a formal criterion of success for the project, at the very least it needs to be measured. As the management aphorism says, "you can't manage what you can't measure." Quantifying the impact of automation: on the customer service experience will at least provide a window into what side-effects the project may be creating, and provide the data to possibly change course and head off disaster. Moving forward without this data is like trying to steer a ship through a fogbank without radar. You think you're making great progress, right up to the moment when you hit an iceberg. Often, no attempt is made to proactively measure the impact on callers of a change to customer service automation, on the mistaken belief that it can't be done until the new system is up and running and taking live customer calls. This myth serves to keep questions of customer impact out of design discussions until the project is complete and it is too late to make changes. Measuring Customer Impact the Wrong Way: Data might not mean what you think it means, or worse, it can be misleading. Data needs to be gathered from representative callers, the measurements have to be meaningful, and the results must be placed in the proper context. A classic mistake is testing a new automated system by using company employees. At first blush, it appears to be an efficient and economical way to get feedback, but company employees are very different from customers . They know how the process is supposed to work, and they know the company jargon. It is also tempting to use "free" statistics from an IVR or ACD to measure the quality of customer service, but many of these measures are at best proxies, and at worst misleading. For example, hold time is an important predictor of customer satisfaction, but it is not the only factor. A call center manager who is offers agents large bonuses if hold times drop may discover unintended consequences, such as agents abruptly disconnecting customers during busy times. 1) Even with good data, it is hard to draw conclusions without context. To reach meaningful conclusions, you need high quality baseline or benchmark data. This can be either data you gathered in the same way at a different time, or benchmark data which is consistent with your own data. Misaligned Incentives for Staff and Vendors: People tend to do what you reward them for doing. A simple rule of management, but one which is often overlooked. If the incentives for the staff and vendors don't match the project goals, then the goals will likely not be met. For example, if improving customer service is a project goal (and it should be), then the project team should have incentives to meet that goal. Just as they would have incentives to complete the project on-time and meet technical performance criteria. Some vendors will balk at contractual obligations to meet customer service improvement goals, but others (those who are more confident in their abilities) will be willing to include this. Regardless of the details, if there's no incentive to meet all the project goals, including service quality, the goals without incentives will inevitably take second place. Focusing on Plumbing, not VUI: A rule of thumb is that 20% of the work in building an automated customer service system will be in the voice user interface (VUI), and the remaining 80% will be systems integration, infrastructure, and other non-interface-related work. It is natural, then, to focus management efforts and planning on the 80% of the project which is plumbing. This is a mistake, since success or failure of the project will almost certainly be determined by the work which goes into designing and implementing the VUI. In a customer service environment, you need to persuade the caller to use the automated options, rather than go to an agent or a competitor. The ultimate success of the project will depend on whether callers actually use the automated system, not on whether it meets the technical requirements. Underestimating the Difficulty of Good Design: Good design of a customer service application isn't always easy, and thinking that it will be simple is almost guaranteed to cause problems. When a customer calls a toll-free phone number, he or she is asked a series of questions, each of which must be answered within a few seconds in order to continue. For many (if not most) callers, this will be the first call in a long time, so there is no chance to train them in how to use the system. A good design must take complete novice callers, and make it possible for them to immediately choose the correct option at each step, as well as provide a way to go back (in case of a mistake), or break out to a human. It should also accommodate the needs of frequent callers to get their tasks done efficiently. In addition, the system should also enhance the brand image of the company, convince the customers that they had a satisfying experience, and ideally, help persuade the customers to buy more products or services. VUI is not GUI: People with experience designing and building Graphical User Interfaces (GUIs) for web sites or desktop software often fail to appreciate the important differences between the GUI world and the VUI world: In a VUI, there are often a lot of infrequent or first-time callers. If you test a VUI by giving more than one or two tasks to the same person, that person is no longer a novice. In a GUI, people often do dozens of different tasks with the same application, and are willing to spend hours learning it. 1)Even a partial VUI implementation is useful, and a good approach to implementing a Big Project. Partial GUI implementations are often useless: what good is a spreadsheet which adds but doesn't multiply? 2) Every VUI usability flaw costs money, directly out-of-pocket from the company, since every problem will cause some people to go to an agent or give up entirely. GUI design flaws don't usually cost money directly, and as long as the software achieves an acceptable level of usability, there's often no benefit to fixing smaller problems. Inflexible Specifications: If I could figure out the magic way to create a flawless specification every time, Bill Gates would no longer be the richest man in the world. Let's face it, every project specification, no matter how good, will need to be changed. This is a fact of life. So treat your vendors as partners in the project's success. Have the flexibility to make changes as they are needed, rather than discovering--after the vendor delivers a good implementation of a flawed specification--that something wasn't right. Assuming Small Changes Will Have Small Impact: "The prototype sounds great, but Marketing wants to make just one little change. You see, they're big on calling customers 'partners,' so we wanted to change the first prompt to say, 'If you're a BigBank partner, press one. If you want to establish a new account, press two.'" That sound you hear is millions of BigBank customers scratching their heads as they wonder if they qualify as "partners" or not. Odds are, their next act will be to dial zero to get to an agent. Of course, it is entirely possible that most BigBank customers are familiar with this odd bit of marketing lingo. It is also possible that most won't have a clue what the computer expects them to do. The only way to know for sure is to test both designs. In the world of customer service, even seemingly small changes can sometimes have an enormous impact. The wrong word in a prompt can send millions of callers rushing to select the wrong option, even though it seemed perfectly reasonable to the people who suggested it. Fragmentation: Fragmentation happens when a company builds its customer service automation out of many small applications, without ensuring that the overall experience is consistent and high-quality. Callers are subjected to a maze of options and prompts, recorded using different voices, and often with no way back after landing in the wrong application. The companies are also poorly-served, since they have no way to know what overall performance is, and the data which might point out problems are hidden across multiple systems. Building a customer service piece-by-piece is a useful strategy. But companies need to be careful to make sure that the overall service and image is consistent with what they want their customers to experience. The whole is not merely the sum of the parts. www.vocalabs.com Contact: Peter leppik pleppik@vocalabs.com

Vocent Announces Industry's First True Multi-Factor Voice Authentication Solution; Vocent DecisionMaker Outperforms Alternative Solutions in Major Pilot Project: Vocent Solutions Inc. announced the industry's first true multi-factor voice authentication solution. The company also released quantifiable results from a large-scale pilot project with Scotiabank, a $200 billion financial services institution. Vocent DecisionMaker is the first voice authentication technology to integrate biometrics, data analytics and content knowledge to make the best possible authentication decision. The accurate, reliable and flexible solution combines Vocent's patent pending technology with speech recognition and voice biometric engines from industry leaders including Nuance and ScanSoft. Vocent DecisionMaker was developed specifically to reduce identity theft and fraud and lower IT costs in the financial services sector and throughout large corporate enterprises. With rampant identity theft costing North American businesses nearly $100 billion annually, Scotiabank conducted a pilot project to evaluate a variety of technologies and methodologies for authenticating customers over the telephone. Included in the extensive pilot was Vocent Confirmed Caller, a packaged application that incorporates Vocent DecisionMaker. The bank piloted Vocent's voice authentication solution as an alternative to content-based approaches, in which agents require customers to answer personal questions or recall personal identification numbers. The content-based approaches compromise security and decrease customer satisfaction by relying on personal information that can be easily discovered. The firm also evaluated authentication systems that use only voice biometrics, but discovered that such systems do not offer strong enough security or sufficient automation rates for large-scale consumer deployments. Vocent's multi-factor system integrates the ease of voice biometrics along with other factors, such as content knowledge and the caller's phone number, with the additional security layers of advanced logic, risk modeling and dynamic dialog management. During the trial, Scotiabank successfully authenticated 95 percent of callers, while reducing false accept rates to below 1 percent. The Vocent system also helps to improve the customer experience by reducing the number of questions callers are asked during identity verification. Many of the most recognized companies, including Marriott, US Bank and Inovant (a Visa Solutions company) employ Vocent's technologies to mitigate risks and improve IT efficiencies. Vocent DecisionMaker is now commercially available as part of Vocent Confirmed Caller and Vocent Password Reset, which can be purchased direct from Vocent or through its expanding sales channel. Vocent is a leader in packaged voice authentication solutions. For the call center, Vocent's solutions offer a tactical way of combating fraud and identity theft while giving customers peace of mind and saving millions of dollars per year. For the help desk, Vocent's solutions automate high-risk requests, such as password resets, eliminating a costly task while maximizing security. At the heart of all of Vocent's products, Vocent DecisionMaker analyzes all available data to ensure the best possible authentication decision. Vocent's packaged applications can be fully deployed in less than 60 days, with seamless interoperability with leading voice platform, speech technology, enterprise software, and call center infrastructure products. Customers that have selected Vocent include US Bank, The Hartford, Cox Communications, Southern Company, First Tennessee Bank, and Inovant (a Visa Solutions company). www.vocent.com or call 650-316-3000. said Rick Davidson, Director, Electronic Banking, Channel R&D at Scotiabank.

One Voice Awarded Speech Recognition Patents in China: One Voice Technologies, Inc. (OTCBB:ONEV)(San Diego), developer of 4th Generation voice solutions for the Telecom and Interactive Multimedia markets, announced that it has received a Notice of Allowance from the patent office in China regarding two of One Voice's patents entitled "Interactive User Interface Using Speech Recognition and Natural Language Processing" and "Network Interactive User Interface Using Speech Recognition and Natural Language Processing". These patents cover human-to-computer speech recognition and Natural Language Processing (NLP) on PC's, embedded and wireless devices. One Voice is now entering the final patent stage in China by filing the appropriate issuance documents and anticipates formal issuance in the coming months. One Voice Technologies, Inc. (OTCBB:ONEV) is a developer of 4th Generation voice solutions for the Telecom and Interactive Multimedia markets. Our Intelligent Voice solutions employ revolutionary, patented technology that allows people to send messages (Email, SMS, Instant Messaging and paging), purchase products, get information and control devices - all by using their voice. www.onev.com Contact: Jim Hadzicki, (858) 552-4466 x116 pr@onev.com

Persay & DTSC sign VAR agreement providing advanced speaker verification solutions for Taiwan’s call center market: Persay Ltd.( Tel Aviv, Israel), a world leader in speaker verification systems for call centers and remote services, and DTSC, a leading provider of call center solutions in Taiwan, announced a partnership agreement for providing advanced biometric speaker verification solutions for Taiwan’s call center market. Call centers are operating under constant pressure to reduce their operational costs. Accomplishing this without compromising on customer service is a great challenge. Speaker verification technology saves call centers operational costs by eliminating cumbersome and inefficient customer authentication procedure, saving precious seconds in each call. Moreover, it allows call centers to introduce new IVR\Speech based service requiring strong authentication. Persay’s speaker verification products VocalPassword and FreeSpeech use the customer's own voice and speech characteristics for authentication. Persay’s products enable secure, efficient and convenient customer authentication within the IVR session or in the background of a conversation with the call center agent. Delhum Technology & Service Corporation (DTSC) is a leading provider of integrated voice/media recording and quality/security monitoring solutions for call centers, financial institutions, telecom operators, government agencies, military and large corporations. The company, founded in 1992, based in Taipei, Taiwan, is a leading supplier of call center solutions. DTSC’s customer base includes most of Taiwan’s banks and telecom operators. Persay is a leading provider of advanced biometric voice verification technology used to reliably verify that speakers are indeed WHO they claim to be. The company's products, FreeSpeech and VocalPassword, enable verification of individuals by voice during natural conversation, or when uttering a simple pass phrase, via any media. Language and accent independent, Persay's products harness the biometric power of voice the same way fingerprints, iris scanning or other physiological and behavioral characteristics are used to verify identity. Persay’s fully-featured products are easily integrated with various IT environments, providing speaker verification services for any application. Persay is a spin-off of Verint Systems (VRNT), a provider of analytic software solutions for communications interception, digital video security and surveillance, and enterprise business intelligence. Persay's headquarters are in Tel Aviv, with offices and representatives worldwide. Contact: Persay Ltd., Shira Lotto, Marcom Manager +972-3-767-8690 Shira.lotto@persay.com www.persay.com DTSC, Alice Chen Marcom Manager +886-2-27124567 Alice.chen@dtsc.com.tw www.dtsc.com.tw Bill Yang, CEO of DTSC

TuVox Sales Surge As Companies Embrace Voice Automated Call Center Solutions: TuVox announced four consecutive quarters of sales growth, pointing to continued momentum on the customer, partnership and technology fronts. New customers such as MCI and Roxio join TuVox's expanding customer roster that already includes industry leaders such as TiVo, Activision and Definity Health. As a result of increased sales and market acceptance, TuVox recently announced the addition of two industry veterans, Terry Shough and Rick Davison, to lead worldwide sales efforts. In the most recent quarter, sales have grown more than 250 percent over the same period in 2003. In addition, TuVox continues to expand its key partnerships with companies including Convergys, VoiceGenie, Verascape, Genesys and Softel, highlighting TuVox's broad acceptance by industry leaders. With TuVox's customizable enterprise speech applications, companies can automate customer service calls, enabling a higher quality of 24x7 service to the consumers, while continuing to offer a live agent option. TuVox's suite of application products illustrates the company's innovative vision to offer an automated "perfect agent" experience, so that customers will freely choose automation over "zeroing out" to a live agent. www.tuvox.com. Contact: Jennifer Balsam SHIFT Communications 415-591-8405 jbalsam@shiftcomm.com

Unveil Technologies Raises Additional $6.5M in Financing : Unveil Technologies, Inc., has raised $6.5 million from investors Sevin Rosen Funds, TD Capital Ventures and Solstice Capital. This latest investment will enable the Company to continue its growth by expanding staffing in all key functional areas, commercialize new technological developments and extend operations into selected international markets. As part of this effort, the Company will be adding to its team of speech and language scientists and developers dedicated to redefining the market for advanced speech self service applications. Unveil Technologies provides voice application software to enterprise call centers that automates customer service interactions. Unlike conventional voice self service systems, Unveil offers a unique approach that blends technology and agents in a collaborative relationship aimed at reducing costs while improving customer satisfaction. Unveil's speech solution, the Unveil Conversation Manager, combines the unlimited capacity of 'virtual’ agents with the knowledge of live agents to enhance the caller experience, increase agent efficiency and automatically improve system performance. Unveil Conversation Manager is a voice self service solution that addresses the most acute problem in today’s call center market - balancing the quality of customer service and the need to reduce costs. Because an estimated 65% of that cost is fixed labor expense, each customer interaction that can be automated or more efficiently handled results in a measurable reduction in operating overhead. However, conventional approaches to call automation like Interactive Voice Response (IVR) and speech recognition can reduce customer satisfaction with confusing touch tone menus or inflexible speech interfaces. In addition, speech systems can be extremely complex and expensive to develop and maintain, and require callers to adapt to highly scripted interactions. Unveil’s Conversation Manager overcomes each of these limitations. The product enables call centers to deploy ‘virtual,’ or software-based customer service agents. Conversation Manager seamlessly blends these virtual agents with live agents throughout a customer dialog to provide a highly-satisfying experience to the caller while significantly decreasing service costs. Rather than simply maximizing call automation, Unveil optimizes each interaction by applying the right mix of resources – speech, live agent, or both – to achieve customer satisfaction and cost reduction. Unveil’s virtual agents: use Unveil’s Meaning Recognition Engine technology to engage callers in flexible, conversational dialogs and automate complex interactions; collaborate with live agents using Unveil’s Conversation Assist technique, increasing labor efficiency and automation rates; dynamically improve over time using Adaptive Learning, an automated process that analyzes real conversations to improve speech recognition performance. Sevin Rosen Funds is a top-tier venture capital firm with a track record of funding successful companies since 1981. The partnership has consistently made early stage investments in pioneering technologies and companies with the potential to create new markets. The most recent fund, Sevin Rosen Fund VIII is capitalized at $600 million. Compaq, Lotus, Cypress, Citrix, CIENA and Capstone Turbine are some of the firm's successful lPOs. www.srfunds.com. TD Capital Ventures invests in early and growth stage companies with significant technology assets and market opportunities that span multiple industry sectors. Specifically, TD Capital Ventures focuses on investments in enterprise software, IT infrastructure, semiconductors, and wireless and emerging technologies. TD Capital Ventures is an investment group of TD Capital. Established more than 30 years ago, TD Capital is the private equity arm of TD Bank Financial Group. TD Capital has approximately $1.9 billion of capital under management through its offices located in Canada and the United States. www.tdcapital.com. Solstice Capital is a seed and early stage venture capital company that invests in alternative energy, education, the environment, life sciences and information technology. Solstice was formed in 1995 and currently has $80 million under management. Representative investments of its partners include Interleaf, Cadance Design Systems, Evergreen Solar Systems, ImaRx Pharmaceuticals, and Maker Communications and Okena. Its general partners are Harry George, Henry Newman, and Fred Bamber working from offices in Tucson, AZ and Boston, MA. www.solcap.com. Contact: Jim Williams, Dir. of Marketing, 781-250-1105 jwilliams@unveil.com

Viztel to raise RM5.25m from IPO: Mesdaq Market-bound Viztel Solutions Bhd plans use proceeds of about RM5.25 million from its listing exercise to fund overseas expansion as well as research and development initiatives. The telecommunications software and voice technology solutions provider started its operations in 1999 and lists China Telecom Corporation and Kuwait Computer Company among its clients. Viztel announced that they would develop a bilingual speech recognition-based telephony system to automate Thai AirAsia's flight schedule and fare enquiries. Viztel's listing exercise involves a public issue of 15 million 10 sen shares at an issue price of 35 sen each. Of the total, 1.5 million shares will be offered to the public, while 11.50 million will be placed out. For the financial year ended Dec 31, 2003, Viztel posted a net profit of RM1.17 million on the back of RM4.61 million revenue. www.viztel.com Contact: Nicole Tan 03 2032 2328 nicole@zi.com.mv chief executive officer Lau Kin Wai

IP Unity Rolls Out Video Chat: IP Unity launched its new Video Chat application at SUPERCOMM 2004. Based on IP Unity's flagship Harmony6000 media server / applications server, this application is the first in a series of video IP-based conferencing and collaboration tools which IP Unity will deliver to service providers for the provisioning of innovative multimedia and 'triple play' services for enterprise and consumer users. The initial Video Chat application is an MPEG-4 based feature set that runs on the applications server IP Unity's carrier customers already have installed in their networks. The new features enable carriers to deliver a bundled service that combines and simultaneously shares live presenter video, PowerPoint slides, stored images and streaming video clips. Combining one-way and two-way communications capabilities, IP Unity's Video Chat will be used in commercial solutions like interactive education, distance learning, corporate broadcasts and multi-point video chat applications. Video Chat is also being adapted for use in simple consumer applications, and is currently in field trial with Asian carriers and consumer subscribers. The product is deployment-ready, shipping to carriers in multiple markets. Video Chat requires minimal hardware and software changes for providers already running Harmony6000-based conferencing and messaging applications. It uses existing audio, video and image processing to gather live and stored content, and formats output to match applications to the end-user device. It enables on-the-fly audio-video conferencing, based on MPEG-4 standards for advanced video and audio encoding, dynamic presentations, and error-free delivery over IP networks. In addition to Video Chat, IP Unity plans to support the following video-rich applications in the coming year. These applications will similarly leverage IP Unity's Harmony6000 and the new generation of MPEG-4 decoders to drive greater productivity, value and revenue opportunities for IP networks: Users will register for video news delivery and receive clips directly in their inboxes as part of unified messaging. Users will register with video news agencies (e.g., cable news stations) to receive real-time feeds. Because they are emails, such items can be viewed / forwarded. No architecture changes are necessary, but a content-serving website is needed for content customization and to facilitate video encoding/transcoding for small-screen devices. Video Streaming delivers live/recorded video transmissions to PC or mobile display devices without full downloading or storage on receiving devices. Streams may be presentations, movies, music, pre-recorded video conferences or clips. No architecture changes are needed to deploy Internet Video Streaming, just video encoding/ transcoding software for small-screen devices. Collaborative conferencing marries audio, data and video conferencing into one integrated experience for multiple participants. A moderator talks and sees other participants, displays slides and charts, or performs collaborative activities (group editing, white boarding, etc). New software delivers the collaborative Graphical User Interface and integrates voice, data and video controls for participants. IP Unity provides IP service solutions for blended TDM and IP networks. The company's Harmony 6000 media/applications server system delivers high speed, volume media processing and carrier-class IP telephony services like audio/web conferencing, messaging, interactive voice response, auto-attendant and prepaid services on a scalable system. 408-582-1100 www.ip-unity.com.

TELUS Deploys IP Unity to Power its IP-One Messaging Services: TELUS (Vancouver, BC), the largest telecommunications carrier in western Canada, and IP Unity (Milpitas, CA) announced that TELUS has deployed IP Unity's Harmony6000 platform to serve as the voicemail and unified messaging engine for its flagship IP-One Telephony Services. In mid-2003, TELUS became the first Canadian telecommunications provider to launch carrier-grade hosted and managed Internet Protocol (IP) Telephony Services. Since then, the company has deployed IP services to business subscribers in the nation's seven most populous provinces. The carrier has selected and deployed IP Unity's Harmony6000 Media Server to upgrade its first generation voice messaging platform to the latest functional requirements and cost performance targets, and to rapidly migrate its voicemail subscriber base to all-IP based messaging by the end of 2004. TELUS (NYSE:TU)(TSX:T)(TSX:T.A) is the largest telecommunications company in Western Canada and the second largest in the country, with more than $7 billion of annual revenue, 4.8 million network access lines and 3.5 million wireless subscribers. The company provides subscribers with a full range of telecommunications products and services including data, voice and wireless services across Canada, utilizing next generation Internet-based technologies. TELUS was a premier founding supporter of the successful bid to bring the 2010 Olympic and Paralympic Winter Games to Canada. www.telus.com. www.ip-unity.com. For TELUS: TELUS Media relations Charles Fleet, 780-493-7227 charlie.fleet@telus.com or For IP Unity: Market Ready Public Relations, Jill Budzynski, Tina Tan, 763-548-8208 jmb@marketready.com Boris Koechlin, Director IP Application Development TELUS

Researchers develop phone-based guidance for the blind: A government-sponsored research project aims to help the blind move freely about town by combining cell phones, wireless Internet, global positioning and voice-recognition technology to tell people where they are. The project, called "Noppa," is being developed by the Technical Research Center of Finland, or VTT, a nonprofit, government-owned research organization. The system is ready for testing this fall. The guidance system, which works with a third-generation cellular phone coupled to a GPS device, tells the users where they are, how to get where they want to go, and gives directions and explains obstacles by voice. "The idea is that the user can say, 'I'm going to such-and-such a place,' and the system responds by saying which bus to take, and how to get to the stop," said Ari Virtanen, a VTT researcher. The device is also hooked up to municipal databases to warn about road and sidewalk construction sites when giving directions, as well as inform about train, streetcar, and bus timetables and possible delays. If multiple buses use the same stop, the Noppa service tells users where their bus is and when it's expected to arrive at the stop so they know which one to board. A news and information service is also available to read users the latest weather reports. The project might include optional applications such as a collision detection system that tells users about immediate obstacles, and a database that could give a spoken version of signs on roads and in the surroundings. The project is going to be evaluated by a group of blind testers in the fall, in Helsinki, the capital of Finland, and the city of Tampere. It's costing about 500,000 euros ($600,000), which is being picked up the Finnish state. The core of the system contains speech-recognition and production software that relays requests and plays back replies in speech — all of which is performed at a central server, not with the device. It identifies street names and basic requests. Operating the device with speech instead of pressing buttons is a crucial feature for the blind, said Juha Sylberg, development manager at the Finnish Federation of the Visually Impaired. "Imagine digging out your cellular phone — which is the size of a large box of matches — in midwinter, wearing thick mittens with rain and sleet whirling around you, and then try to start keying away at these tiny buttons," Sylberg said. Although the project is not commercial, VTT's research often serves as a basis for marketable products. And in this case, the researchers believe the features and the products they're developing and testing would have a lot of appeal to a general population. Indeed, some of the same technologies are already combined in other products, such as car navigation systems. Tourists in San Francisco can rent little buggies known as GoCars that combine GPS with an automated voice that serves as a tour guide. "A number of similar devices are already out on the market," Virtanen said about the Finland project. "What we're doing instead is trying to estimate the range and variety of services that can be provided, and how reliable they are in real life."

FTW Builds Multimodal Mobile Applications Using Kirusa's Multimodal Solution: Forschungszentrum Telekommunikation Wien (FTW) (Vienna, Austria) the premier telecommunications research institute in Austria, and Kirusa (Edison, NJ) announced that FTW will use Kirusa's Multimodal Solutions in its research program on mobile multimodal user interfaces. The research program, funded by leading carriers and information technology providers in Austria, involves building and showcasing several multimodal applications for delivery on a wide range of mobile devices on 2.5G and 3G wireless networks. FTW's MONA project (Mobile multimOdal Next generation Applications, http://mona.ftw.at) is funded by Kapsch CarrierCom AG, Mobilkom Austria AG and Siemens Österreich AG together with the Austrian competence centre programme Kplus. The project is showcasing the value of multimodality in a messaging application, called MONA@Work, and in a multi user real-time game, MONA@Play. Consumers can entertain themselves on their phones with a "Who Wants to be a Millionaire" style MONA@Play, while enterprise customers can manage email, voicemail, SMS, and MMS with MONA@Work. Enabling wireless carriers, service providers, and enterprises to offer applications with integrated voice and visual interfaces, Kirusa's multimodal solutions allow mobile users to simultaneously or alternatively use voice, text, graphics, keypad, and stylus to interface with wireless services and applications. The solution integrates seamlessly with industry standard technologies such as SMS, WAP, XHTML, VoiceXML, and HTTP, and supports all major mobile networks, including CDMA, CDMA 1x/3x, GSM, GPRS, UMTS, Flash OFDM, WiFi, and others. Enabling compelling multimodal end-user applications on a wide variety of handsets from all major manufacturers -- carriers and enterprises offering this dynamic service are able to monetize voice and data applications immediately, leveraging the new smart devices, as well as devices already widely deployed in their networks. Kirusa's Circuit Voice multimodal platform, IP Voice Multimodal Platform, and IP Voice Multimodal Clients have been integrated into FTW's MONA architecture. Forschungszentrum Telekommunikation Wien (ftw., Telecommunications Research Center Vienna) is the joint research center of the leading players in business and science responsible for developing telecommunications in Austria. It brings together departments from Vienna University of Technology, several well-known telecommunications companies, some smaller enterprises, the Austrian Research Center Seibersdorf and the Association of the Electrotechnical and Electronics Industry. The expertise accumulated at ftw. makes it the primary address for Austrian telecom companies, as well as one of the leading institutions worldwide for application-oriented research in telecommunications. www.ftw.at. Kirusa, Inc. is a multimodal infrastructure company that has developed industry leading multimodal technology (with several patents pending) for the delivery of data applications using multimodal interfaces. Kirusa's customers include France Telecom, Bouygues Telecom, and mobilkom austria. +1 (732) 623-2118 or visit www.kirusa.com Contacts: Inderpal Singh Mumick Kirusa, Inc. Tel.: +1 732 623-2118 ext. 217 Email: mumick@kirusa.com . Georg Niklfeld, project manager of the MONA project at FTW.

Mobiles to recognize Indian languages: OnMobile (Bangalore), a mobile solutions provider, has launched its Automatic Speech Recognizer (ASR) for Indian languages, the first complete and comprehensive language model developed and deployed for the Indian market. This landmark development ushers in the capability for Indian consumers to interact with automated services with simple voice commands in their native languages. The new model currently supports major Indian languages like Hindi, Marathi, Gujarathi, Bengali, and Punjabi. Support for Tamil, Telegu, Kannada and Malayalam will be available shortly. OnMobile is deploying the Indian English language model for commercial service with the leading wireless operators in India, including Airtel, Hutchinson, Idea Cellular and Tata Indicom. Over 8 million unique cellular subscribers all over India use this technology to enjoy OnMobile-powered services such as Cricket, RingTones, Musicards, Jokes, Contests, CheckMate, Liquidate, etc. Following the successful deployment of these speech-enabled multimodal Value Added Services, OnMobile has extended the power of speech recognition to call center automation for telecom operators and financial services institutions. A wide range of M-Commerce and enterprise applications, including wireless vertical market solutions are also being launched. Development of the speech recognition model for Indian languages was a challenge. In the project spanning over 9 months, OnMobile worked closely with the Central Institute of Indian Languages (CIIL) – Mysore and Nuance (NASD: NUAN). Several hundred thousand speech samples were collected that included words, phrases, sentences, numbers and connected digits. Linguists from CIIL provided their expertise for the pronunciation styles and their variations across the Indian sub-continent. The exercise also involved transcribing and tagging all the speech samples, which were finally then processed to build the Speech Recognition Model. The speech samples were collected through different access channels like landline, CDMA and GSM phones to account for the acoustic impact that these communication modes have on the speech recognition technology. OnMobile's Multimodal systems are being deployed for Value Added Services by Mobile and Fixed-line Service Providers in India. These systems currently handle over 4 million premium airtime calls every month. www.onmobile.com Arvind Rao, CEO of OnMobile Systems, Mouli Raman, CTO of OnMobile +91-80-2668 0182

ScanSoft Introduces OpenSpeech Attendant: ScanSoft, Inc. (Nasdaq: SSFT) announced the availability of OpenSpeech Attendant 1.0, the company's standards-based auto-attendant solution. Based on the SpeechWorks Division of ScanSoft's SpeechAttendant 8.1 architecture, OpenSpeech Attendant offers partners and developers a packaged, VXML-based application to strengthen their portfolio of speech application assets for both new and existing customers. Standards-based platform providers using OpenSpeech Attendant leverage a dictionary available anywhere -- containing well over a million names. As a result, intensive tuning time typically associated with the deployment of speech-enabled auto-attendants is decreased dramatically. Additionally, OpenSpeech Attendant maximizes the performance and accuracy of OpenSpeech Recognizer speech recognition and natural-sounding Speechify text-to-speech software. With improved accuracy, simplified system administration and support for large directories containing more than 100,000 names, OpenSpeech Attendant is a packaged application offering. By extending its auto-attendant offering to its standards-based platform partners, enterprise organizations worldwide can optimize their existing investments in a VoiceXML infrastructure. OpenSpeech Attendant is immediately available. www.scansoft.com/speech.

ScanSoft Releases SpeechAttendant 8.1, ScanSoft, Inc. (Nasdaq: SSFT) announced the availability of SpeechAttendant 8.1, its turnkey auto-attendant solution. The latest release features improved support for large name directories and additional system administration functionality in ScanSoft's effort to further ease system management and decrease the total cost of ownership. SpeechAttendant has more than 1,000 installations worldwide and handling over one-half billion calls annually. The new SpeechAttendant release draws on caller and customer feedback from more than 1,000 installations worldwide. It provides efficient, personalized service to greet and route callers based on the number dialed and divert internal and inbound calls away from live operators so that they are free to focus on more value-add service. Unlike touch-tone systems, SpeechAttendant easily handles complex menu choices without forcing the caller to listen to a long list of options and eliminates the frustrating dial-by-name experience. This automatically reduces "zero outs" and caller frustration, resulting in shorter call transfer and hold times for callers and reduced telephony costs for the enterprise. SpeechAttendant customers typically realize a return on their investment in less than one year. SpeechAttendant 8.1 incorporates the performance and accuracy of OpenSpeech Recognizer speech recognition and natural-sounding Speechify text-to-speech software, the SpeechWorks Division of ScanSoft's field-proven speech technologies. It allows callers to use natural phrases such as, "May I speak with Martha Jones." Enhanced features and additional capabilities include the following: *Improved Out-of-the-Box Accuracy for Wireless and Noisy Environments SpeechAttendant 8.1 incorporates technologies from OpenSpeech Recognizer software to precisely discriminate between background noise and speech and dramatically increase accuracy rates for wireless, hands-free and noisy environments. It also features SpeechWorks' patented LEARN capability that automatically adapts based on caller usage, to ensure the highest possible accuracy in any application. *Improved Support for Large Directories SpeechAttendant 8.1 delivers enhanced performance for large directories, containing more than 100,000 names, with improved disambiguation techniques to offer the same, high quality caller experience found with smaller directories containing fewer than 1,000 names. * Eased System Administration SpeechAttendant includes several added features to improve system performance and provide a satisfying caller experience without significant support from the system administrator. It includes an extensive nicknames template that automatically associates common nicknames to proper names, such as adding -- Bob, Bobby, Rob to Robert -- eliminating manual entries by the system administrator. It also provides a large department acronyms template that automatically imports department acronyms present in directory source files and converts them into formats that can be used by text-to-speech and speech recognition engines, such as converting "mktg" to "marketing" or "svcs" to "services." SpeechAttendant also now supports the HL7 protocol required by healthcare providers and hospitals to speech-enable access to patient rooms without relying on operators and touch-tone systems. In addition, SpeechAttendant 8.1 adds new system performance reports to its extensive list of reports. *Improved Fault-Tolerance and Support for Complex Multi-Site Deployments SpeechAttendant further enhances its redundancy and multi-site capabilities by allowing customers to easily accommodate robust real-time redundancy and load-balancing. In addition, SpeechAttendant 8.1 offers increased flexibility for deployments in complex, distributed architectures as speech-enabled auto-attendants become mission-critical systems. SpeechAttendant is distributed throughout North America and Europe through channels such as Bell Canada, British Telecom, NACR, NEC BNS, NextiraOne, SBC, Telus and others. The SpeechWorks division has broadened its channel distribution by adding new distributors including Call Processing, eNabling Technologies, Servion, SSP Telecom, The Via Group, and Vitec. In addition, the SpeechWorks division now offers a VoiceXML-based version so that standards-based platform providers worldwide can take advantage of the market leading auto-attendant solution. SpeechAttendant is available immediately. www.scansoft.com/speech.

UbiCall Virtual Receptionist at Belgium Ministry of Finance: UbiCall Communications has been selected, together with IBM and Cisco to supply the Ubicall Virtual Receptionist Software, which uses the Nuance Speech recognition, to the Ministry of Finance here in Belgium. The Belgium Ministry of Finance, will be moving to a new building in the coming few months. They will install a totally new VoIP Telecom infrastructure in this new Building. IBM together with Cisco have been awarded this major contract, which is regarded as one of the largest VoIP deals so far, and will certainly give a big "push", towards the general acceptance of VoIP. UbiCall Communications, as a subcontractor to IBM, has been awarded with the supply of the ASR powered Virtual Receptionist called Voxplorer, as an integrated part of this new VoIP infrastructure. Voxplorer is implemented in the Cisco Callmanager 4. This solution will be put into service before the end of September 2004, initially for about 4000 users. The system is ultimately dimensioned for about 30.000 users (= the whole ministry), and will be expanded gradually to accommodate the full 30.000 users. UbiCall Communications SA/NV is a Belgium-based technology company that specializes in developing interactive telephony solutions. UbiCall was founded in March 2000 and has analyzed the interactive telephony market, with a specific focus on the emergence of voice recognition technology within the sector. UbiCall has acquired considerable insight into the state of current technology offerings and existing automated telephony solutions and concluded that there are considerable drawbacks to existing systems, particularly with respect to the implementation of speech recognition and has seized the opportunity to apply a different approach to combine speech recognition and automated features, providing user-friendly, cost-effective, and efficient telephony solutions. UbiCall offers the VoxPlorer Receptionist on either an ASP or a CPE basis. UbiCall partners include: Cisco, Alcatel, Avaya (Switching partners) and with Genysis, Envox, Voxpilot as platform vendors. Customers include: Belgacom, Dexia Bank, Ethias NRB, Saint-Peter Hospital Brussels, Delhaize LME Business Center University of Fribourg, Opel/Saab Motor Willems, and SBMI. Contact: Willy Van de Voorde, Director of Sales, Ubicall Communications sa/nv Mobile: (+ 32) (0) 475/ 24 44 90 Office Tel: (+32) (0) 65/32 15 46 email: willy.vandevoorde@ubicall.com www.ubicall.com/

Nortel Networks, VoltDelta Announce Venture to Focus on Directory and Operator Services: Nortel Networks (NT) (CA:NT) and VoltDelta (New York), a wholly owned subsidiary of Volt Information Sciences, Inc. (VOL) have entered into an agreement under which Nortel Networks will contribute certain assets and liabilities of its directory and operator services (DOS) business to VoltDelta. In return, Nortel Networks will receive a minority equity interest in VoltDelta, a leading provider of DOS in North America and Europe. This transaction between the companies is expected to close during the third quarter of calendar year 2004. Nortel Networks expects approximately 160 DOS employees in North America and Mexico to join VoltDelta. Nortel Networks will retain its Traffic Operator Position System (TOPS) business, based primarily in Research Triangle Park, N.C. VoltDelta's DOS business will provide continued focus and investment to develop next generation directory assistance voice over packet based solutions - particularly for the wireless market - to meet rapidly evolving customer requirements. The venture will provide existing and future DOS customers with new solutions, giving customers access to an expanded suite of products, content and enhanced services. VoltDelta, a subsidiary of Volt Information Sciences (reported as Volt's Computer Systems segment) is a leading provider of enhanced directory assistance solutions and information services to the global telecommunications market. Thirty years of operator services experience have enabled VoltDelta to meet evolving market requirements providing innovative technology and services to wireline and wireless markets www.voltdelta.com or www.volt.com Nortel Networks is an industry leader and innovator focused on transforming how the world communicates and exchanges information. The Company is supplying its service provider and enterprise customers with communications technology and infrastructure to enable value-added IP data, voice and multimedia services spanning Wireless Networks, Wireline Networks, Enterprise Networks, and Optical Networks. As a global company, Nortel Networks does business in more than 150 countries. www.nortelnetworks.com Nortel Networks. Joe DiAngelo, president, VoltDelta, Sue Spradley, president, Wireline Networks, Nortel Networks Contact: Cecelia Pfeiffer, Director of Marketing cpfeiffer@voltdelta.com

Phonetic Systems And V-Enable Launch Two Voice Activated Multimodal Consumer Applications with Major Wireless Carrier: Phonetic Systems (Billerica, MA) and V-Enable (San Diego) announced that US Cellular has launched two multimodal applications for their customers. SAY IT. BIZFINDER and SAY IT. TONEFINDER allow callers to use simple voice commands, instead of having to triple tap buttons to enter text, in order to download information, directions and ring tones. SAY IT. BIZFINDER, is a multimodal directory assistance application that allows callers to search, find, and locate business listings nationwide by simply speaking their input. It does what calling Directory Assistance can do, but it can do a lot more, like retrieve driving directions, maps, placing a call, etc. SAY IT. TONEFINDER allows users to search, select and download ring tones using their voice. For example, now users can speak the name of the artist or song title, and retrieve the ring tones that match their input, instead of having to type in the text on a small keypad. The combined applications received over twenty thousand hits in just the first week of availability. The technology behind these applications is Phonetic Systems' Voice Search Engine, a highly dynamic and scalable speech recognition engine, and V-Enable's veANYWAY multimodal solution which enables services that combine the flexibility and richness of data services with the ease-of-use of speech commands. The partnership's simultaneous multimodal solution provides for an exciting end-user experience which can be applied to numerous applications. SAY IT. BIZFINDER and SAY IT. ONEFINDER are currently available on BREW enabled phones. Phonetic Systems delivers global speech recognition solutions that provide voice access to information, people, applications and services. The company's patented Voice Search Engine (VSE) is designed specifically for voice searching very large directories or any size database with frequently changing, untuned datasets of information. The VSE can power packaged applications or it can be integrated into any platform or environment to deliver speech applications like carrier Directory Assistance, voice access to CRM applications, password reset, automated attendant and more. Phonetic Systems' US corporate headquarters are in Billerica, Massachusetts, with global offices in the United Kingdom and Israel. www.PhoneticSystems.com V-Enable is a leader in developing, delivering and enabling carrier infrastructure software that brings a rich user experience for today's mobile user, based on the Company's patented veANYWAY Multimodal technology developed using X+V standards. Multimodal technology promises to accelerate the adoption of data and speech services for carriers worldwide. Though offering standards-based solutions, V-Enable provides the carrier with maximum flexibility and extensibility of their network. The company is headquartered in San Diego, CA with sales and engineering offices in London, Munich, Paris, and Delhi, India. www.v-enable.com Contacts: Heather Howland, 978.439.3600 Phonetic Systems hhowland@phoneticsystems.com Jim Benedict V-Enable Corporation 858.481.8366 jim.benedict@v-enable.com

InfoTalk-Recognizer 8.2 integrated with Aculab: The latest version of InfoTalk's conversational speech recognition engine, InfoTalk-Recognizer 8.2, has been successfully integrated with Aculab's Prosody media processing resource platform. With the Prosody interface, InfoTalk-Recognizer 8.2 is loaded with more advanced features, including automatic channel adaptation, noise cancellation and multiple model activation, that further enhance speech recognition accuracy. InfoTalk-Recognizer 8.2 has also introduced a user-friendly Application Programming Interface (API) and architecture for rapid development of voice enrollment applications. www.infotalkcorp.com. www.aculab.com Contact: Dr. Alex Leung, CEO of InfoTalk. info@infotalkcorp.com or +852 2190 9600 Katie Chaplin, Aculab Telephone: +44 (0) 1908 273 800 Email: katie.chaplin@aculab.com

Audium Successfully Completes Interoperability Testing with Cisco: Audium (New York) announced that Audium 3 met the Cisco AVVID (Architecture for Voice and Video Integrated Data) Partner Program test criteria for interoperability with the Cisco Internet Service Node (ISN), version 2.1. The complete Audium package is a standards-based software solution that offers the fastest application development, the easiest way to add speech recognition to IVR, and the best way to connect callers to the information they need. In use at the largest enterprises and service bureaus in the world, more VoiceXML applications have been created and deployed worldwide with Audium than any other application package. Audium joins the Cisco AVVID Partner Program as a Contact Center participant. Through participation in the program, Audium provides a complete voice application development package that will help users of Cisco ISN launch more VoiceXML applications that provide better automated self service by sharing information with existing enterprise systems. The Cisco AVVID Partner Program sets criteria for interoperability testing and helps leading product and services firms to deploy innovative business solutions. The program provides enterprise customers with information regarding Cisco AVVID partner products and services that the Cisco AVVID partner believes meet the program's criteria for interoperability with Cisco networking technology. Cisco Customer Contact products include solutions for IP contact centers, intelligent contact management within a time-division multiplexing (TDM) environment and self-service applications such as interactive voice response (IVR). These solutions enable customers to move into the next phase of customer contact—a Customer Interaction Network, a distributed, IP-based intelligent information network that provides enhanced responsiveness, streamlines customer interactions, and extends customer service capabilities across the entire organization, giving businesses a more integrated and collaborative approach to customer satisfaction. Audium enables IVR teams to create phone applications that provide better customer service by making it easier to integrate ISN with enterprise business applications and web systems. By using open standards, contact center customer service applications built with Audium and Cisco ISN can leverage information and transactional technology in place today such as Remedy ARS, PeopleSoft, SAP, Siebel, and Oracle. Audium creates VoiceXML application software that revolutionizes customer service IVR systems. Audium's core product, +1.877.AUDIUM.1 or info@audiumcorp.com Contacts: Tom Stephenson Audium Tel: +1.212.609.4000 E-mail: tstephenson@audiumcorp.com . Barry O’Sullivan, VP and GM of the Cisco Customer Contact Business Unit.

ClickFox Customer Monitoring Solution: ClickFox, Inc. announced the availability of a Customer Behavior Intelligence solution suite designed to model, measure and improve self-service applications. This solution allows businesses and government organizations to translate complex customer interactions across a variety of self-service applications including, interactive voice response (IVR) systems, speech recognition, Web applications, customer relationship management (CRM), interactive kiosks and mobile devices -- into fact-based decisions that optimize single-channel and cross-channel business performance. Using ClickFox's Customer Behavior Intelligence suite, organizations are able to analyze, understand and optimize their customer interactions within their self-service applications and can dramatically reduce operational costs, improve customer satisfaction and revenue generation and enhance the overall interactive customer experience. ClickFox's Customer Behavior Intelligence suite is based on VIBE (Visualized Interactive Behavioral Engine), ClickFox's patent-pending technology in the area of user behavior modeling and optimization. ClickFox's founding team includes several faculty members and researchers from Georgia Tech who have applied their advanced academic research into software modeling solutions that enable companies to visualize and map user interactions across a variety of interactive systems. The ClickFox Customer Behavior Intelligence suite is a comprehensive decision support platform that includes system modeling, flow and behavior visualization, root cause analysis and reporting and an advanced recommendations engine. The suite can be applied to a variety of IVR/speech, Web and cross-channel self-service applications. The Customer Behavior Intelligence suite consists of the following: * ClickFox for IVR/Speech * ClickFox for Web * ClickFox for Cross-Channel. ClickFox's Customer Behavior Intelligence suite provides a comprehensive array of modeling, analysis and optimization features and functions for optimizing self-service applications: Behavior and Application Visualization -- Allows organizations to visually see traffic, task and transaction flows for any step or task in self-service applications. ClickFox provides a unique interactive visual system capability that is intuitive and easy-to-use for both business and technical analysts. It enables companies to find problems areas and get essential information on user conversion rates through analyzing user tasks or interaction sequences. This helps organizations pin-point drop-off points at any point in the customer transaction process and follow the "failures" so that business-critical scenarios can be optimized. Cross-Channel Modeling and Analysis -- Using ClickFox's business modeling capability, the entire customer self-service experience can be optimized. Customer behavior can be viewed and analyzed across multiple systems, including IVR, speech, Web and CRM systems. Organizations are now able to see the entire experience in one system view -- allowing organizations to understand and improve customer experience across channels. ClickFox analyzes aggregate customer behavior and can link behavior to individual utterances in speech applications or to recorded phone calls to identify and fix the highest impact customer service issues. Automated Optimization Engine -- ClickFox's optimization engine utilizes artificial intelligence to enable organizations to identify immediately those application areas that are misaligned with user expectations and behaviors. This engine compares application usage with system structure and mathematically generates robust recommendations for improvement based upon customer behavior. Recommendations deliver specific actions to help streamline call flows or restructure interactive flows. Pricing and Availability The ClickFox Customer Behavior Intelligence suite is value-priced according to call or visitor volumes and target application(s). Customers receive telephone support, maintenance, warranty and updates during subscription term. ClickFox Customer Behavior Intelligence suite is available now from ClickFox or from an authorized ClickFox alliance partner. ClickFox's software modeling solutions enable its customers to translate complex customer interactions across a multiple self-service channels -- such as interactive voice response (IVR), speech recognition, Web sites, customer relationship management (CRM) systems, kiosks and mobile devices -- into fact-based decisions that optimize single channel and cross-channel business performance. Privately held with leading investors that include Cedar Fund, Delta Ventures and Vertias Venture Partners, www.clickfox.com 404.351.8020. Tal Cohen, president and CEO of ClickFox

Edify announces developer Network and donates auto attendant speech application to support voice and speech community developers gain access to Edify Virtual Attendant source code, tools and Open Dialogue: Edify Corporation (Santa Clara), an S1 subsidiary, announced the launch of the Edify Developer Network and the availability of an open-source version of the Edify Virtual Attendant (EVA). The Edify Developer Network allows developers to take advantage of the latest tools and open-source applications available and share creative ideas to shorten development cycles and improve system functionality. Through the Developer Network, developers gain access to information and applications enabling them to reduce development time and costs. In addition to developer discussion forums, the Edify Developer Network includes an application sharing area with real-world examples, helpful tutorials, sample code and other resources to assist community members with complex development projects. Additional development tools are available in the download area including, documentation, patches, updates, utilities and more to support the development of applications using Edify and its supported third-party vendors. Within its Developer Network, Edify also provides free access to EVA, the open-source version of its virtual auto attendant. As a prepackaged application built on top of the Edify platform, EVA provides automation of one of the most repetitive tasks at any organization - transferring calls to the appropriate party. EVA's powerful speech interface creates a conversation where the caller simply asks for the person or department they wish to speak to. EVA enables companies to transfer callers to the right department or geographic location without spending time in a queue, resulting in quicker interactions, accurate information and improved customer satisfaction. The Edify Developer Network is located at http://developer.edify.com and is available to Edify direct customers and partners. With fourteen years of industry leadership and innovation in voice and speech applications, Edify has changed the way more than 2,000 global companies enhance customer service. 408-982-2000 (USA), 81-3-5532-7827 (Japan) or +44 162 885 0222 (Europe, Middle East, Africa) www.edify.com www.s1.com Contact: Edify Corporation Cynthia Dimidik, Edify Corporation: (408) 486-1791; cynthia.dimidik@edify.com Heidi Rosenberg Nadel Phelan, Inc. Strategic * Technology * Communications Phone: 831-440-2405

Speech Enabled SIP VoIP VoiceXML Platform Integrates with Asterisk, an open source PBX: Invores Systems, Inc (Long Island, NY). announces integration of its SIP VoIP VoiceXML gateway platform, EXpress2.0, with Asterisk, an open source PBX. For over thirty years enterprises have relied on automated self-service solutions to meet customer service needs and contain costs. As labor and benefits costs rise, these systems continue to be critically important to an overall bottom line. Enterprise customers and service providers deploying the Asterisk PBX will be able to easily enhance their offerings with either speech enabled or touch tone self-service applications. One example of such an EXpress2.0 application is Invores’ First Impressions, a speech enabled virtual receptionist system that runs on EXpress2.0. Asterisk has a touch tone based auto attendant feature, but for service providers and companies looking to differentiate, First Impressions simulates a live receptionist. Callers can just say the name of the person or department they are calling. EXpress2.0 runs in a Windows environment, with a VoiceXML2.0 execution environment and uses speech recognition and synthesis software from Scansoft SpeechWorks to simulate "real" conversations. Invores is a privately held Long Island, New York based software development firm focused on Interactive Voice Web speech-enabled customer self-service solutions in commercial, service provider and government organizations. www.invores.com 516.868.5455. Tom Multer, President of Invores

Pronexus Introduces 'Get-Up to Speech': Pronexus Inc.( Ottawa, Canada) announced the availability of Get-Up to Speech, a rapid application developer bundle for Microsoft Speech Server. Available for a limited time from Pronexus or its distribution partner CYGCOM, this special package lets developers rapidly create, test and deploy sophisticated Microsoft Speech Server applications in VB.NET and many other languages, all for an unmatched price of only US$1695. Speech applications have proven to deliver customer satisfaction, cost savings and operational efficiency in many different industries. Get-Up to Speech provides IT Managers, software programmers and call center developers with all the components necessary to experience the power of speech for themselves. 'Get-Up to Speech' consists of the following components: ·Microsoft Speech Server 2004 Standard Edition, 180-day Evaluation Edition: Microsoft Speech Server 2004 combines Web technologies, speech processing services, and telephony capabilities into a single, integrated product. Speech Server lets companies unify their Web and telephony infrastructure and extend ASP.NET applications for speech-enabled access from a variety of devices. ·Pronexus VBSALT 1.0: VBSALT is the first and only rapid application development tool for Microsoft Speech Server. It provides developers with a graphical design environment for call flows, encapsulates best practices for speech application design, reduces the learning curve for developers new to speech and/or telephony and lets programmers leverage existing programming skills. Runtime licenses are included in the bundle. ·Intel Dialogic D/41JCT-LS Combined Media Board: This four-port, analog, converged communications board is fully compatible with Microsoft Speech Server and supports voice, fax, and software-based speech recognition processing in a single PCI slot. Includes Intel Dialogic Continuous Speech Processing Technology, a set of firmware features optimized to support high-performance, speech-enabled applications. ·Intel NetMerge Call Manager: This telephony interface manager (TIM) is a software module that provides fast and easy integration of the Microsoft Speech Platform with the D/41JCT-LS telephony board. ·Intel® Dialogic System Release Software: This software package includes all necessary drivers for the telephony boards as well as a suite of installation, configuration, and diagnostic utilities that can further simplify telephony server development. As a Microsoft Speech Partner and Intel Communications Alliance Member with more than a decade of experience in telephony & speech, Pronexus provides a full range of technical support offerings. For the Microsoft Speech Server, a dedicated SALT division delivers custom application development, best practice consultation, code review services and more. Additional professional services and training are also available. Founded in 1993, Pronexus Inc. has over 3,500 customers worldwide and is the developer of VBVoice, and the new VBSALT, Rapid Application Development tools for business-critical telephony and speech solutions. Pronexus also provides VeoSuite, a range of turnkey speech applications for Microsoft Speech Server and a complement of professional services that voice-enable a variety of industries and applications. Comprehensive support and acclaimed training complete the firm's offerings. CYGCOM is a leading Microsoft Speech Distribution Partner and an authorized Value Added Distributor of Intel products. CYGCOM distributes speech and telephony integration products from the industry's most reputable manufacturers. CYGCOM offers solutions that are complete, powerful, fully programmable, scalable, and capable of running mission-critical enterprise communications applications. Through it's US and Canadian facilities, CYGCOM provides products, consulting services, on-site technical support, eSupport, helpdesk, systems integrations, education and marketing services for channel customers. www.cygcom.com Contact: Sonia Langenberg (Tercler), Marketing Communications, Pronexus Inc., Tel: (613) 271 8989 ext. 541 sonia.langenberg@pronexus.com Jeff Valliant, President & CEO of CYGCOM

SandCherry and LumenVox Partner to Offer Speech Solutions: SandCherry, Inc (Boulder, Colo.), and LumenVox, LLC. (San Diego, Calif.) announced a partnership to offer premier-quality speech and touch-tone solutions to service providers and enterprises. SandCherry will offer LumenVox’s Speech Recognition Engine (SRE) and LV Speech Tuner with its AppDev VXML development toolkit, AppPackage pre-integrated application solutions, and custom solutions using the SoftServer speech services platform. The combined solution can support a wide range of applications for delivery to phones and multimodal devices. Customers have the choice of creating independent voice-enabled applications, or extending existing web-based applications by adding a Voice User Interface for a fraction of the cost of standalone applications. Enterprises can extend the reach of customer service, sales force automation, supply chain management, and employee portal applications by using the phone to increase availability, productivity, and user satisfaction. Service providers can realize a completely new set of services by combining the value of data with the ease of use offered by speech. LumenVox, is a speech recognition company with over a decade of telephony experience. They develop a suite of speech recognition software that includes the Speech Recognition Engine (SRE), Speech Driven Information System (SDIS), LV Speech Tuner and ArtiSpeech. Download trial versions at: www.lumenvox.com/. www.sandcherry.com/

SandCherry Selects Paraxip Technologies’ Media Gateway to Simplify Telephony Connectivity: SandCherry, Inc., has selected Paraxip Technologies’ speech-optimized media gateway software to simplify telephony connectivity into its SoftServer platform. By using SandCherry’s Telephony Connector powered by Paraxip, companies will have a cost-effective, pre-integrated, and certified solution for connecting the SoftServer platform to a wide variety of telephone switches, PBXs, and ACDs. The Paraxip software provides the SoftServer platform with an integrated media gateway that has the capability to convert phone calls on analog and digital lines or trunks to Session Initiation Protocol (SIP) standard Voice over IP (VoIP). The SandCherry Telephony Connector powered by Paraxip runs on standard, off-the-shelf telephony and media processing hardware and offers comprehensive connectivity features such as wide protocol support, advanced transfer capabilities and speech-optimized media processing capabilities. The SandCherry Telephony Connector powered by Paraxip is available for trial now and is scheduled for general availability in August 2004. www.sandcherry.com/. Paraxip Technologies offers software products enabling a seamless transition from traditional telephony to IP telephony. Paraxip’s products are designed from the ground up to offer Voice-over-IP (VoIP) applications with un-paralleled transparency to the Public Switched Telephone Network (PSTN), enabling seamless operations even in hybrid environments, where traditional and IP telephony equipment and applications are mixed. www.paraxip.com. Contact: Paraxip Technologies, Serge Forest, Phone: (514) 286-1171 x 217, serge.forest@paraxip.comSal Barcia, Vendeloo Inc. 212-475-3738 sal@vendeloo.com

ScanSoft Releases OpenSpeech DialogModules 2.0: ScanSoft, Inc. (Nasdaq: SSFT) announced the availability of OpenSpeech DialogModules 2.0 software, ScanSoft's patented application building blocks used to accelerate application development and deployment using a consistent user interface to ensure the highest levels of caller success in speech service transactions. The SpeechWorks Division of ScanSoft pioneered and patented DialogModules in 1996, and since that time, SpeechWorks and its partners have deployed more than 750 speech application using DialogModule technology. OpenSpeech DialogModules have been adopted by industry-leading, standards-based platform providers and tool vendors to accelerate application development of custom speech applications as well as support SpeechPAKs, SpeechWorks industry specific packaged speech applications. OpenSpeech DialogModules free software developers from programming frequently occurring speech recognition tasks in new speech applications. When used in conjunction with OpenSpeech Recognizer 3.0, OpenSpeech DialogModules provide leading recognition accuracy, efficiency, and caller satisfaction on the industry's broadest selection of telephony platforms. Included among the OpenSpeech DialogModules are capabilities for collecting dates, time, yes/no responses and addresses. For example, rather than writing code to collect a correct spelling for a name, developers can use the Name OpenSpeech DialogModules to accelerate application development while providing a consistent user interface for the caller. The latest release includes a number of powerful application development enhancements including: *Latest Standards Support: OpenSpeech DialogModules are components that generate VoiceXML and provide an application building block for simplifying the development of VoiceXML applications. The latest release supports multiple drafts of the VoiceXML 2.0 specification including the latest candidate recommendation from February 2003. *Lower Cost of Ownership: OpenSpeech DialogModules provide improved caching performance resulting in greater CPU savings and faster response time for callers. For example, OpenSpeech DialogModules 2.0 can now support more than 400 cacheable channels, a four fold increase in performance relative to OpenSpeech DialogModules 1.2. As a result, speech solutions that employ OpenSpeech DialogModules 2.0 software can lower web application server infrastructure costs by 75 percent. * Improved Application Management: OpenSpeech DialogModules 2.0 improve application management and tuning by providing for server side logging of information. Application developers can quickly and easily access call logs making it easier to deploy or optimize their speech applications. *Supports Development of Complex Prompts: OpenSpeech DialogModules 2.0 provide greater flexibility for user interface designers enabling them to dynamically change parallel grammars, properties and behaviors. For example, designers can employ text-to-speech software to spell back information using the Name and Address OpenSpeech DialogModules, incorporating the same voice used for the pre-recorded prompts, to create a seamless caller experience. * Wide Application Server Support: OpenSpeech DialogModules 2.0 support a wide range of J2EE application servers supporting the J2EE Servlet 2.3 and JSP 1.2 specification independent of the operating system. * OpenSpeech Xccelerator Testing Suite: OpenSpeech DialogModules 2.0 includes a testing package for platform partners. OpenSpeech Xccelerator includes code, grammars, audio prompts, documentation, analysis tools and metrics so that platform partners can quickly qualify and integrate OpenSpeech DialogModules on their platform. www.scansoft.com . Contact: Marie Ruzzo, ScanSoft, Inc., Tel: (617) 428-4444 marie.ruzzo@scansoft.com

Excel, Vocalocity Partner on VOIP: Excel Switching Corp. (Hyannis, MA), a provider of carrier-class open services platforms, media gateways and media servers for communications networks, and Vocalocity (Atlanta, GA), a provider of VoiceXML and SALT software designed for OEM and VAR partners, announced a partnership to deliver a high-performance, low-cost VoiceXML platform designed explicitly for service provider deployments. The offering will make it easier and significantly less expensive for fixed and mobile providers across VoIP, traditional TDM and converged networks to provide touch-tone and speech-enabled applications on a broad scale. The joint solution provides direct integration between the Excel AnyGen Converged Services Platform (CSP) and VocalOS from Vocalocity. VocalOS provides pre-integration with the Vocalocity VoiceXML interpreter, several leading speech recognition and text-to-speech engines, and the Vocalocity App Center GUI-based VoiceXML development environment. This initiative will enable application developers to seamlessly combine Excel's superior call control and media processing capabilities with VoiceXML dialogs and speech recognition, making it easier for developers to offer advanced interactive voice services on a single proven platform. The VocalOS will also be backward compatible with existing Excel SwitchKit applications. www.excelswitching.com www.vocalocity.net Contacts: Excel Switching Corp. Bill Kelly, Director, Marketing Programs, 508-862-3113 bkelly@xl.com Vocalocity John Hibel VP Marketing and Business Development 404-487-1200 ext. 1381 jhibel@vocalocity.net Mike Twomey, vice president, product planning for Excel mtwomey@xl.com

ART Will Supply smARTspeak XGT For LX5550 Voice-Controlled Phone From LG Electronics: ART Advanced Recognition Technologies announced that LG Electronics has selected the smARTspeak XGT embedded speech interface for its latest voice-controlled CDMA phone, the compact LX5550 clamshell. The Korean electronics and communications company is the world's largest producer of CDMA phones. The LX5550 is the second CDMA phone from LG to incorporate the new smARTspeak XGT. The smARTspeak XGT is fully speaker independent, meaning it can be operated by any user "right out of the box" - with no voice training of the device. It also adds a new voice feedback capability that gives users audio confirmation of spoken names, numbers, and other commands. This is done by "reading" from the phone's 499-name contact list using ART's advanced text-to-speech technology. LX5550 Features Accurate Voice Control "Right Out of the Box" The features of the smARTspeak XGT will enable users of the LX5550 phone to dial by speaking a person's name or phone number. Giving equally fast and accurate response whether users say "John Smith" or "Dial 1202111002", smARTspeak XGT maintains ART's out of-the-box, multi-user experience. In addition, the phone will recognize commands such as "contact" and "redial" and offers trainable digit dialing that can be adjusted to a user's accent. When more than one name is available, ART's "Best Match" function presents a short-list of options both as text and voice for extra convenience. In addition to its speaker-independent voice recognition capability, the new LX5550 allows users to access contacts, digit dial, redial, retrieve voice mail, and time all by voice command. The smARTspeak technology also makes the new LX5550 one of the safest mobile phones available. By enabling total phone operation by voice, smARTspeak XGT is ideal for in-car use. With its built-in speakerphone, it offers a hands-free mode for improved driving safety without the need or expense of a special car kit. ART's embedded software-only solutions are deployed in dozens of product lines from industry leaders such as Motorola, LGE, Xelibri by Siemens, Panasonic, Mitsubishi, NEC, BenQ, Pogo, Logitech, AlphaCell, Capitel, Compal, Inventec, Casio, RAKS, Siemens Automotive, PaceBlade, Hitachi, and others. ART has strategic and technological partnerships with: Texas Instruments, Qualcomm CDMA Technologies, Motorola Semiconductors, Intel, Infineon, Agere, ADI, TTPCom, DSPG, ParthusCeva, Microsoft, Symbian, UIQ, SVOX, Accelent Systems, Phone-Or, and others. www.artcomp.com Contact: ART Advanced Recognition Technologies Ltd.Betty Finkelstein, +972-3-7685018 betty@artcomp.com or ART Advanced Recognition Technologies, Inc. Gabi Artzi, 805-581-3999 gabia@artcomp.com Jae-Gyung Ham, vice president of CDMA Marketing at LG.

Three of Top Four Indianapolis 500 Finishers Leverage Clarity Technologies' CVC(R) Noise Canceling Software Inside Motorola XTS 5000(TM) Two-way Radio to Improve Critical Communications: Clarity Technologies Inc. (Troy, MI), announced that its Clear Voice Capture (CVC) technology, embedded into Motorola XTS 5000 two-way wireless radios, improved the quality and performance of voice communications for the second, third and fourth place winners of this year's Indianapolis 500. This enhancement demonstrates that the combination of these two technologies delivers a solution to voice quality and performance degradation issues in the most challenging of audio environments. The solution improved audio communications between the drivers, crew chief and spotter radios. The crew was able to clearly hear the driver as the award-winning voice technology worked to eliminate the engine noise. Clarity's CVC technology inside both the crew chief's and spotters' radios also reduced noise in the pit lane, enabling the driver to understand critical messages without the need for repetition or confusion. One crewmember even reported that the radio worked so well during the race that he thought the engine was turned off while the car had already begun moving around the track. This also was the first time an XTS 5000 digital radio was used at the Indy 500. Usually, the digital mode is used during the practice sessions to improve range and privacy. This latest generation mission critical radio proved to be the best mode of communications at the track. Cellular networks on race day were very overloaded with phone calls and text messaging, but the Motorola mission critical two-way radio solution continued to effectively work through the entire rain-delayed event. Clarity Technologies Inc., a venture-backed company established in April 1998, develops and markets a suite of software products and algorithms (CVC - Clear Voice Capture) that dramatically improve wireless voice communications, the quality and accuracy of voice-based user interfaces and speech recognition software programs, particularly in noisy real world mobile environments. Clarity is the only associate member of the Bluetooth SIG that provides noise and echo cancellation technology. ClarityLabs is a services division of Clarity Technologies Inc. and provides companies with the ability to evaluate select and integrate speech recognition technology for their voice-based applications in addition to providing porting services and soon an independent and objective voice quality certification process. Motorola, Inc. (NYSE: MOT - News) is a global leader in wireless, broadband and automotive communications technologies that help make life smarter, safer, simpler, synchronized and fun. Sales in 2003 were U.S. $27.1 billion. Motorola creates innovative technological solutions that benefit people at home, at work and on the move. The company also is a progressive corporate citizen dedicated to operating ethically, protecting the environment and supporting the communities in which it does business. www.claritytechinc.com www.motorola.com Contact Tom Houy, VP Marketing (tomh@claritytechinc.com ) or 248-822-5122.

Clarity Technologies Inc. awarded additional patents for Superior Audio Voice Quality in Extremely Demanding Environments Clarity Technologies Inc. (Troy, MI) has received its third patent in the area of improving the quality and performance of voice communications. This patent deals with the integration and processing of a sensing device including an array of sensor processor cells arranged into a detection array for a single device. Clarity currently has approximately 20 additional patents pending which include technology for Voice Signal Extraction, Directional Sound Acquisition, Audio Visual Speech Recognition, Noise Reduction in Subbanded Speech Signals and a Multi-Microphone System for a Handheld Device. The combination of these technologies delivers solutions able to solve voice quality and performance degradation issues in the most challenging of audio environments. Clarity Technologies' first patent was issued in September 2003 for Blind Signal Separation which is a signal processing system with at least one sensor detecting a number of input signals including at least one source signal, a signal separation architecture that defines a relationship between the input and output signals, an update processor coupled to the architecture for determining a rate of change for time varying parameters, and an output processor coupled to the architecture for providing output signals. In February, a second patent for Adaptive State Space Signal Separation was granted which is a signal processing system and method for separating a number of input signals into a number of output signals. www.claritytechinc.com Contact: Tom Houy, VP Marketing tomh@claritytechinc.com or 248-822-5122

VoiceSignal Provides Multiple Language Speech-activation for Samsung A790: VoiceSignal (Woburn, Mass) announced the first implementation of its new speech application. This new VoiceSignal application provides support for English and Spanish speakers on Verizon Wireless’ Samsung SCH-A790 wireless enterprise phone. The SCH-A790 global phone works on both GSM and CDMA networks, enabling Verizon Wireless enterprise customers to stay connected and access data from a single device throughout the world. By incorporating VoiceSignal Voice-Activated Phonebook and other voice features into the SCH-A790, Verizon Wireless is meeting the growing needs of business users for interfaces that make wireless devices more usable and their features more accessible in any environment. In addition to being the first phone with both English and Spanish speaker-independent voice recognition, the Samsung SCH-A790 features the following enhancements to VoiceSignal’s suite of voice features: Improved interface – The SCH-A790 embodies an innovative new speech user interface that incorporates animated cues that indicate to the user when to speak and what they can say. In addition the voice prompts make it easy to use the phone in an "eyes-free" mode, so the user does not have to look at the display in order to look up a name, dial a number or launch an application. Embedded User Guide – The SCH-A790 is the first phone to contain VoiceSignal's user guide, which is available by pressing a soft key labeled "Help" from various parts of the application. The embedded context-specific user guide answers common questions about the application and its features. New Personalization Options - With the SCH-A790, users can customize the voice interface to satisfy their personal preferences. In the Voice Settings menu, users can adjust: o Voice prompts – Users can turn voice prompts on and off. With voice prompts turned on, the system employs VoiceSignal's user-friendly voice prompts to guide the user through a task. By turning voice prompts off, expert users can bypass the prompts while issuing commands for a faster dialing experience. Recognition results – Users can determine whether or not they would like to confirm recognition results prior to dialing as well as set the application to either display or not display choice lists. o Recognition matching - Users can set the recognizer "sensitivity" to adjust recognition matching to optimize the user experience. Other VoiceSignal features on the SCH-A790 include: Digit Dialing - Allows users to dial any number, including those not in the contact list, by speaking the numbers, with no speech training required. Voice-Activated Phonebook – users can look up or dial any name in their device’s phonebook, regardless of the number of names. Being speaker-independent, the VoiceSignal technology allows users to easily dial numbers and navigate the handset’s menus, without first having to "train" the device. Text-to-Speech Output – by combining speech recognition and text-to-speech output, the VoiceSignal application enables users to control dialing features without having to use their hands or look at the display, making it easier and more enjoyable for consumers to use the a610. Command and Control – enables the user to launch applications on the handset by speaking a single command such as "Launch Browser" which opens the browser on the phone or "Launch Voice Memo" which allows the user to leave a memo. VoiceSignal is a leading supplier ofvoice interfaces for mobile devices. Users can speak a name, a number, or a command, and get connected in one step. VoiceSignal’s unique architecture is compatible with the entire range of mobile devices – from entry-level voice-only handsets to enterprise-class PDAs and smart phones. VoiceSignal is shipping on millions of units manufactured by the largest global handset brands and sold by major operators worldwide. www.voicesignal.com Contact: Mira Genser, VoiceSignal, +1 (617) 513 7020 mgenser@voicesignal.com

Text Messaging Use in U.S. Could Explode As Voice Systems Grow, Improve: On an average day, customers of Verizon Wireless, the biggest US cellphone company, use their handsets to send and receive more than 23 million short text messages. As big as that number might sound, US message traffic has reached only a fraction of the levels in text-crazed areas of Europe and East Asia, where 2 billion text messages a month are the norm for some countries, such as the United Kingdom. But in many ways, it's perplexing that anyone would use wireless text messaging. After all, typing 55566688833 to spell the word love would strike a lot of people as less than practical. Even with software that can reduce the number of keystrokes required to spell a word on a cellphone, there's no escaping the fact that the 12-character keypad developed by Bell Labs engineers in the 1960s for making phone calls remains much less convenient for spelling words than a regular keyboard. By the end of this year, however, a start-up company called Voice Signal Technologies Inc. expects to be selling speech-recognition software for cellphones that would let people compose and transmit short messages (to a maximum of 160 characters) just by dictating them aloud and saying "send." Voice Signal, whose technology already powers dial-by-name features in several Samsung handsets, says it can support speech-to-text conversion of up to 30,000 words through a program that would take just 1 megabyte of memory in the 16- or 32-megabyte chip that provides the brains inside many cellphones. Once sending a message like "What's up? Want to meet at Mary Anne's for a beer at 7?" becomes as simple as simply saying the words, it's easy to imagine that text messaging could explode. The implications could be even more enormous for markets such as China, where "writing" a single Chinese character can require entering five to 10 digits. Mandarin and Cantonese are two of the 14 languages for which Voice Signal has developed speech-to-text systems. The impact that effective speech recognition would have on text messaging, however, is only the tip of the iceberg. If it proved to be reliable, the technology could replace the need for keypads and screens on phones, because users could say a phone number instead of dialing it. The "user interface" aspects of a handset -- particularly full-color display screens -- generally cost far more than the phone's innards, things like microprocessors and digital radio chip sets. Once you can make a phone that doesn't need a keypad or screen, you can dream of everything from cellphones embedded in jewelry and pens to super-cheap screenless units that could make wireless communications affordable for billions of poor residents of the Third World. "With full voice recognition, you could make phones very, very cheaply," said Richard J. Geruson, a former senior vice president with the Finnish cellphone giant Nokia who became Voice Signal's chief executive last year. By the end of this year, Voice Signal expects Samsung and other handset makers, including Motorola, Panasonic, and a contract manufacturer for Audiovox, to produce 20 million units with its speech-recognition technology built in. Currently, the main application lets people place calls without dialing, typically by holding down the star key to get to the speech prompt, then saying something like "Call Ellen at the office" or "Call Ellen on her mobile phone." Coming soon are systems that will let people navigate through often-bewildering menus for downloading games or ring tones or pulling up weather forecasts. Also, as an intermediate step, Voice Signal has developed a "phrase recognition" system that would let people compose reasonably nuanced text messages by stitching together a series of canned phrases: "Hi, it's me." "I got the information you sent." "Thanks for sending it to me." "I'm out of the office now." "You can reach me on my cellphone." "Talk to you soon." To go to the full open-ended speech dictation system, new users would have to "enroll" by speaking a string of 120 words that appear in succession on the phone screen, so the system recognizes the speech patterns of the person using the phone. The process, which starts with "Because, papers, teaching, community, especially" and continues through all the pronouns and prepositions and other words, takes about three minutes. We trained a prototype Samsung phone with the Voice Signal system last week, and found it imperfect, but surprisingly accurate. You have to speak slowly and deliberately, but it was able to accurately render even weird messages such as "I am eating the dog food, and it tastes pretty good." Arriving at the cusp of being able to have phones that you not only talk on but talk to has taken nine years of effort by a phalanx of Voice Signal's PhDs. Rather than some silver-bullet breakthrough, "It's a thousand little things that got us here," said cofounder and president Dan Roth. "Speech technology has been disappointing people for 20 years," he said. "But the biggest challenge for us right now is educating people, letting people know that these capabilities are available. We're nowhere near the end of the road in terms of what we can do." www.voicesignal.com

Spok21 speaking out loud for those who can't: Igel, a German company specialised for the last 20 years in accessibility, has chosen Acapela text to speech to voice enable Spok21, its innovative communication device designed for people with speech disorders. Spok21, which has been recently unveiled to the market, is a transportable communication tool for disabled people looking for an easy-to-use, ergonomic product packaged in a colourful and well designed device. Spok21 has embedded Acapela text to speech in order to read out loud any typed information, in real time, with a natural sounding result, opening up a new way of communicating. Designed with everyday use in mind, Spok21 aims to be a convivial and smart companion to help users communicate with other people. With the large display and ergonomic keypad, the users can easily type in their message. The text to speech result is rendered loud and clear via two integrated loudspeakers for optimal intelligibility. A large memory size allows the users to store their own recurrent messages. Six shortcut keys provide a quick access to short phrases. A long battery life ensures autonomy and last but not least, design has been pushed up to the colour touch with a choice of four different colours. Spok21 is available in 5 languages and already marketed in Germany, Netherlands, Belgium, Switzerland, Austria, Sweden and UK. www.rehavista.de and on IGEL at www.igel.rehavista.de . www.acapela_group.com Contact: Caroline Houël chouel@acapela-group +33 (0)5 62 2471 22

Loquendo Technology Powers LARTEC Home-Domotic Solutions: Loquendo, the global speech technology company, and LARTEC, Spain's leading player in Home Domotics development - signed a Value Added Reseller Agreement whereby Loquendo's high-quality multilingual Text-To Speech will be offered within LARTEC's state-of-the-art solutions. Households throughout Spain will enjoy Jorge and Jordi new Loquendo's male personas for the Spanish language. Within the context of this agreement Lartec is releasing the first home automation system, which talks in any language and enables one of its principal attractions: the ability to communicate interactively with users. Lartec's home automation systems allow users to control their homes in an easy and intuitive manner. Thanks to a 15" colour TFT tactile screen equipped with stereo speakers, the system manages single home elements centrally, as well as accessing a whole range of services on the web. Easy home environment management is now a reality with Lartec's systems. Loquendo's technology enables the home automation system to talk to users, informing them about their home environment, and even alerting them by phone should any incident occur during their absence. Loquendo's speech technology approach delivers the scalability and reliability required to deploy voice applications of any complexity. Featuring a comprehensive, multi-lingual set of technological components such as its multilingual Text-To-Speech software (Loquendo TTS), Loquendo offers a realistic, "lifelike voice" for the dynamic data and prompts required in today's voice applications. With over 30 years R&D expertise, Loquendo is at the forefront of the global speech technologies market. Its high-quality, high-performance technologies (Loquendo TTS text to speech engine, Loquendo Embedded TTS text to speech engine, Loquendo ASR automatic speech recognition) and platforms guarantee the best solutions in 15 languages and 33 voices - with more coming soon. From its headquarters in Turin, Italy, Loquendo deploys technologies that power over 2,000,000 calls every day in the telecommunications and enterprise markets throughout the world.www.loquendo.com The national enterprise Lartec possesses delegations in the Spanish cities of Valencia, Madrid and Barcelona. Lartec is planning for this year the opening of new delegations in Marbella, Valladolid, and Bilbao. It has to be emphasised that the new entrance of the Grupo Inova Capital and other minority investors from the total investment group, which until now belonged to Obradis group, actually leads to consolidate Lartec as the major home automation capital enterprise of the country.

Virtual Speaker gives a voice to iProgress' leading e-learning programme PROGESSION: iProgress, an active e-learning company with over 400 clients, designs, develops and installs training and complementary e-learning solutions for professional users of computer applications. iProgress has chosen VirtualSpeaker, Acapela Group's virtual recording studio solution, to voice-enable their integrated e-learning solution, PROGRESSION, already licenced to 3 million users worldwide. The PROGRESSION solution provides training contents, either available for standard office applications, or developed specifically for Business Applications (ERP, CRM, specific sector applications... ). The PROGRESSION contents now integrate sound, thanks to Acapela's vocal solution, and also multimedia illustrations and software simulations for easier learning. By using VirtualSpeaker, iProgress completes their training software with a smooth and natural voice, in several languages, with no costly recording and updating expenses. Acapela's text to speech, which can transform any text into speech in real time with an extremely natural voice, provides quality high enough to replace a human voice recording. The localisation of the contents into different languages can be done with the one text-to-speech tool, and the updates are much easier to manage. VirtualSpeaker frees you from the ties of recording studios and the associated logistics issues. Personalising information, increasing learning efficiency, saving time, and flexibility are the main features that make speech synthesis an interesting component for convivial and interactive e-learning. www.iprogress.com . www.acapela_group.com Contact: Caroline Houël chouel@acapela-group +33 (0)5 62 2471 22 Jocelyne Dereeper, Director of Instructional Design at iProgess

Talking heads help hearing-impaired take phonecalls: New software being developed may give the deaf and hard-of-hearing a new way of receiving phonecalls – using a computer generated face that users can lip-read. The SYNFACE software connects to regular computer microphone and soundcard hardware, which in turn is connected to the telephone line. The software can make and receive calls in the usual way. When a connection is made, a face appears on screen which syncs in time with the speaker on the other line. The EU-funded project, which has been in development for the past three years, is part of the new UCL Centre for Human Communication. The centre opens today and is intended to bring together academics and researchers from many different disciplines. Dr Andrew Faulkner, who heads the project, said that the software is a unique development. "The recognition software hears vowel and constant sounds and replicates them on screen. It's very different from most speech recognition software, which uses a lot of contextual information. It's a very distinct method; operating over a very short timeframe of speech, and it has to be pretty much in real time. There's only a delay of 150ms." Dr Faulkner also agreed that the software might dramatically improve disabled access to NHS phone-based services, such as NHS Direct. Penaran Higgs from RNID: "Deaf and hard-of-hearing people face many technological barriers in their daily lives that can be greatly improved with the help of new developments such as this. Evaluations currently being run by RNID aim to gain measures of the effectiveness and usability of SYNFACE. So far the results have been very encouraging, with exciting implications for the future." SYNFACE is currently in final tests. For a video demonstration of SYNFACE: www.phon.ucl.ac.uk/home/andyf/SynfaceHIGH.wmv

Kurzweil Educational Systems Announces Kurzweil 1000 Version 9: Kurzweil Educational Systems, Inc., announced Kurzweil 1000 Version 9, the premier scanning and reading software for people who are blind or visually impaired. The latest version of the software makes printed pages and electronic archives more accessible and usable than ever before with a new online encyclopedia search and retrieval capability, broadened support for Web repositories, more accurate scanning, enhanced DAISY and portable device support, and better navigation through document linking. Kurzweil 1000 Version 9 new features give users: *Expanded Online Search and Retrieval - Support for online encyclopedias, such as Columbia Encyclopedia and Wikipedia from Wikimedia Foundation as well as access to additional Web repositories, including American Foundation for the Blind’s (AFB) AccessWorld and Talking Newspaper Association of the United Kingdom, complement the online book and magazine search capabilities and the already notable list of accessible Web repositories. *More Accurate Scanning - A new Optical Character Recognition (OCR) engine, ScanSoft OCR, and an update of the ABBYY FineReaderTM OCR engine provide the most accurate recognition available. *Additional Study Tools - Notes and Document Link options expand an already powerful set of study tools. The Notes feature provides an additional way to annotate documents. The new Document Link feature allows users to explore existing links to dynamically navigate within the same document, another document or Web site to access needed information. *Increased Portability - Support for Book Port extends the list of portable devices to which Kurzweil 1000 can transfer text and MP3 files. *Talking Business Calculator - A new full-function talking business calculator performs, for example, financial functions for loans, mortgages, leases, savings, annuities and more. Version 9’s new features expand upon the extensive list of existing features in Kurzweil 1000, including state-of-the-art synthetic speech options, color scanning, flexible writing and editing tools, the ability to send files to e-mail, Braille embosser compatibility, and photocopier, fax and calculator applications. Kurzweil 1000 Version 9 is scheduled to ship to customers later this summer and can be previewed at the 2004 National Federation of the Blind (NFB) Convention, June 29-July 5 in Atlanta, Georgia; the 2004 American Council of the Blind (ACB) National Convention, July 3-10 in Birmingham, Alabama; and the 2004 Association for Education and Rehabilitation of the Blind and Visually Impaired (AER) International Conference, July 13-18 in Orlando, Florida. A single license of Kurzweil 1000 retails for $995. Volume pricing is also available. Through October 29, 2004, Kurzweil Educational Systems is also offering customers: *A discount of $100 off the retail price of any new, full-product purchase of a Kurzweil 1000 license. *The opportunity to participate in the new Refer a Friend Program in which registered Kurzweil 1000 users who refer a new customer who buys a Kurzweil 1000 license receive a $50 American Express gift certificate. Kurzweil 1000 can be purchased directly from Kurzweil Educational Systems at 800-894-5374 (within the U.S. or Canada) or 781-276-0600 (outside the U.S. or Canada), or from Kurzweil Educational Systems' worldwide network of resellers listed at http://www.kurzweiledu.com/resellers.asp. www.kurzweiledu.com. Contact: Collaborative Insight, Inc. Ellenann Haughey 617-357-7715 haughey@collaborative.com

Medquist Uses Handheld and PDA Dictation Technology: Physicians can now have accurate patient information and secure dictation capabilities in the palm of their hands. Keep track of patient schedules, manage daily workflow and record dictation with a Palm OS or Pocket PC handheld device. Busy physicians who practice at multiple sites can easily monitor schedule changes from home or the office. Patient information is synchronized to a handheld device using wireless technology or a card reader attached to a PC. Dictation can be uploaded any time, anywhere with accurate patient demographics attached, ensuring accuracy and accelerating the transcription process. Handheld/PDA solutions are easy to install and require minimal technical support. MedQuist’s products, services and technologies offer the right blend of flexible health information management solutions including an Internet-hosted workflow management and typing platform, coding services, dictation systems, document distribution, speech recognition, template-based products and transcription services. The PhysAssist IQ solution is easy to install, requires minimal technical support and facilitates remote management of client and handheld software. MedQuist offers flexible health information management solutions through the right blend of products, services and technologies, including Internet-hosted technology, coding services, digital dictation systems, document distribution, speech recognition, template-based products and transcription services. www.medquist.com 800-624-5681 x3216 solutions@medquist.com

ChartLogic, Inc. Selects ScanSoft's Dragon NaturallySpeaking SDK for Its Comprehensive Electronic Medical Records - EMR – Solution: ScanSoft, Inc. (Nasdaq:SSFT), announced that ChartLogic, Inc, the leader in voice-driven electronic medical records (EMR) has selected ScanSoft's Dragon NaturallySpeaking SDK, the leading speech recognition software integration solution, for its flagship EMR solution, ChartLogic. ChartLogic utilizes voice recognition not only for dictation, but to drive the entire electronic charting process. Working closely with physicians, ChartLogic leveraged the Dragon NaturallySpeaking 7 SDK to create a comprehensive, voice-directed electronic medical chart. ChartLogic's solution provides physicians with an accurate, easy-to-use alternative to paper charts that enables physicians to save time and money - resulting in more time to focus on patient care. Traditional charts, requiring handwriting, dictation or manual transcription, are time consuming and expensive, and are often not completed by the physician until the end of the examination day. With ChartLogic, physicians can complete their charts in less than 60 seconds at the end of each patient visit - resulting in substantial time savings and more accurate patient notes. By using the speed and accuracy of voice-driven dictation and commands, ChartLogic also enables physicians to more easily and completely document their services for billing purposes. ChartLogic has developed a comprehensive package of easy-to-use software and hardware, known as Precision Voice technology, to optimize its voice capabilities for the clinical environment. Precision Voice technology was designed in conjunction with physicians to mirror the normal processes of a physician practice. ScanSoft's Dragon NaturallySpeaking 7 SDK technology portfolio provides healthcare providers, technology suppliers and integrators with the ability to quickly and easily speech-enable commercial and in-house PC- and Web-based Healthcare Information Systems (HIS) and Electronic Medical Records (EMR) applications, and delivers unrivaled speed and accuracy. Through their use of the Dragon NaturallySpeaking SDK, healthcare organizations can save thousands of dollars per doctor each year in reduced or eliminated manual transcription costs. www.ScanSoft.com ChartLogic, Inc. was founded in 1994 and has since created the most innovative and comprehensive voice driven Electronic Medical Record system available today. ChartLogic, Inc. has integrated clinical experience and technological expertise to create a system that is uniquely designed for clinicians and their diverse environments. www.ChartLogic.com . Jack Savage, CEO for ChartLogic www.scansoft.com . Contact: Marie Ruzzo, ScanSoft, Inc., Tel: (617) 428-4444 marie.ruzzo@scansoft.com

Philips releases first Hungarian speech recognition software for Radiologists: Royal Philips Electronics (Vienna, Austria) announced the release of the Hungarian Radiology ConText for its speech recognition software SpeechMagic. Philips is the first company to take speech recognition for the medical profession into Hungary. The development of the ConText was carried out in co-operation with Medisoft Software Development and Services Kft. ("Medisoft") – a provider of medical information technology for the Hungarian market. The ConText incorporates a Hungarian user interface and the general interface to application systems and is available from Medisoft. SpeechMagic can be integrated into any Radiology Information System (RIS) or Hospital Information System (HIS) available on the Hungarian market. www.philips.com Contact: Philips Speech Processing, Speech Recognition Systems, Valerie Gray, Tel: +43 1 60101 1542 valerie.gray@philips.com www.speechrecognition.philips.com Ágnes Nyers Dr., managing director of Medisoft.

Philips SpeechMagic boosts transcription productivity in Aiken Regional Medical Center: Royal Philips Electronics(Vienna, Austria) announced that Aiken Regional Medical Center in Aiken, S.C. significantly increased the productivity of report transcription by implementing the Philips speech recognition software SpeechMagic, integrated into the WinScribe digital dictation system. Aiken Regional Medical Center is a 225-bed acute care facility. Like many healthcare facilities, Aiken was looking for a solution to gain control of its Radiology transcription backlog and reduce operating costs. Therefore Aiken implemented a digital dictation system with integrated speech recognition. Aiken staff use WinScribe Internet Author with the SpeechMike and a barcode scanner to dictate reports. Once a report is dictated and recorded, the audio file is immediately routed to the SpeechMagic server for conversion from speech to text. The transcriptionists receive both the audio file and the recognized text within minutes of dictation. They then correct recognition errors and finalize the report. Subsequently the completed report is automatically merged into the Radiology Management System (RMS) and filed in the corresponding patient record. Reports are available to physicians on average, 90 minutes after the radiologist dictates the report. In addition to the operational benefits, the patient service has improved. Treatment can start shortly after examination, because long waiting times have been eliminated.. Contact: Philips Speech Processing, Speech Recognition Systems Valerie Gray Tel: +43 1 60101 1542 valerie.gray@philips.com www.speechrecognition.philips.com so2say communications Armin Scheuer Tel.: +49 30 219 13 914 armin.scheuer@so2say.biz Howard Sanford, Aiken Radiology Department Manager .

ViewPlus to Incorporate Text-To-Speech Technology from Wizzard Software in New Chameleon Product Line: Wizzard Software (OTCBB: WIZD) announced that its text-to-speech technology offering has been selected by ViewPlus for inclusion in their new Chameleon product line. Chameleon offers a way for the visually impaired to interact directly with graphical information using a proprietary touch pad and software. Instead of requiring sighted assistants to alter information and images to accommodate Braille labeling, visually impaired computer users can simply "print" an image and using Chameleon, touch areas of the graphic and while feeling the "visual" information, have text and other characteristics read out-loud through computer generated speech. The end users of ViewPlus products are mainly blind students in the K-12 and College/University level but the Company also serves a growing population of corporate and government clients with visually impaired employees. There are approximately 93,600 visually impaired or blind students and at least 1.5 million blind and visually impaired Americans using computers today. What drew ViewPlus to Wizzard Software was Wizzard's ability to provide best-in-class synthetic voice technology and cater to their special needs. Professor John Gardner, a blind physicist from Oregon State University, set out to found ViewPlus eight years ago in Corvallis, OR. After a dramatic turn of events left Dr. Gardner blind in mid-career, he was presented with a challenge: comfortably retire or develop the technology necessary to remain a prominent fixture of the academic community. With his naturally positive spirit and inquisitive mind, Dr. Gardner chose the latter, and ViewPlus was born. Dr. Gardner's research and development ranges from innovative hardware and software to revolutionary new ways of using Braille for math and science access. In the five years ViewPlus has offered commercial products, it has doubled sales every year - going from a basement lab at Oregon State University to a 30+ employee corporation serving customers Although ViewPlus distributes a myriad of software and hardware products like the Tiger Software Suite and the world's only Accessible Graphing Calculator, Dr. Gardner's premier hardware development is the ViewPlus embosser series. What makes the ViewPlus embossers series like no other Braille printer in the world is they are all Powered by Tiger, ViewPlus'proprietary embosser technology. This allows ViewPlus embossers to interact with Windows software just like ink printers - creating a perfect tactile copy of exactly the same information - text, graphics, charts, etc. - just like an ink printer. The embossers even use different dot heights as colors get darker or lighter, just like an ink printer would use more or less ink. ViewPlus offers the only embossers with this seamless compatibili ty with mainstream Windows software. www.wizzard.com www.viewplus.com 541.754.4002 info@viewplus.com

Vocollect Expands Channel Program to meet demands for Voice-directed Distribution Systems: Vocollect announced the expansion of its channel program by naming Speech Interface Design, Inc. (SID), a Master Value Added Reseller (VAR). The move comes in response to a significant and continuing increase in demand from a wide variety of mid-tier companies seeking the well documented benefits of the integrated Vocollect voice-directed system. SID, a voice systems integrator with more than 10 years of experience installing and supporting Vocollect's voice systems, will work with specialty market systems integrators in North America to deliver Vocollect's offering to their customers. The Vocollect integrated voice-directed distribution solution combines high-performance industrial speech recognition technology with its rugged, built-for-purpose, full-shift Talkman wearable, mobile computer and a complete suite of Vocollect's distribution dialogs for warehouse functions including picking, receiving, put-way, transfers, replenishment, and loading. Until recently, Vocollect's channel program in North America, which has generated 1300% revenue growth in the past three years and will exceed 50% of Vocollect's North American revenues this year, was concentrated in a select set of top-tier warehouse management systems (WMS) providers and material handling equipment (MHE) systems integrators. Founded in 1993, SID recently marked its tenth year as an authorized Vocollect value added reseller. Co-owned and managed by Ahmed Reza, a former Vocollect employee, SID has developed and installed numerous Vocollect voice solutions for customers in a variety of vertical markets. Among those companies working with Speech Interface Design to integrate and resell the Talkman voice system with their existing solution offerings are Softeon, a warehouse management system (WMS) provider, in Dulles, Va.; and ASAP Automation, a turn-key systems integrator based in Louisville, Ky. Speech Interface Design, Inc., was founded in Pittsburgh, Pa., in 1993. The company provides fully customized, turnkey and fully integrated voice distribution systems, which include needs analysis; initial system design; programming development; integration and testing; system installation; end-user training and technical support. SID is an exclusive Vocollect-certified supplier and installer of the Talkman system. 412-323-1135 www.speech-interface.com Since pioneering voice recognition for industry in 1987, Vocollect has set the standard for warehouses and other business settings. Together, its market-leading Talkman wearable mobile computer and integrated software suite cut operating costs by eliminating errors and improving worker productivity shift after shift. Leading WMS providers, material handling integrators and specialty voice solution providers around the world choose the performance and reliability of Vocollect to deliver integrated voice-directed distribution solutions. And every day, over 60,000 workers around the globe rely on Vocollect's integrated voice solution to help improve their operations. www.vocollect.com North America: info@vocollect.com Europe: voc_emea@vocollect.com Margaret Ransom Vice President of Communications Desbrow & Associates mransom@desbrow.com Phone: (412) 471-5700 . David Kogan, vice president of ASAP Automation

Loquendo Releases Loquendo ASR 6.3 Technology: Loquendo, the global speech technology company, released Loquendo ASR 6.3 software technology, which includes important application development enhancements and it is available in two new languages: Mexican and Portuguese. Loquendo ASR is the next-generation speech recognition technology for speech-enabled applications. It is speaker-independent and reliably recognizes large-scale vocabulary continuous speech, even in the noisiest environments, including wireless. Loquendo ASR's knowledge is based on a comprehensive set of reusable and expandable Recognition Objects, which are combined with a development toolkit for effective and user-friendly recognition capability packaging The latest release includes a number of application development enhancements including: * the "Language Pack" concept, allowing a clear distinction between the core engine and the specific language knowledge: now different languages can be upgraded and installed separately. * a new powerful semantics software, providing the flexibility to easily support the new and changing standards. * the compilation time has been greatly reduced, especially in case of complex grammars. * the Loquendo ASR neural engine has been furtherly optimized, running now on integer arithmetic: this means a significant reduction of computational time without degradation of recognition accuracy. Loquendo ASR makes it possible to create user-friendly services that are highly complex in terms of vocabulary size and interaction flexibility. Loquendo ASR is available in a wide range of languages: Italian, Spanish, Catalan, Portuguese, U.S. and U.K. English, French, German, Greek, Brazilian, Argentinean, Chilean, Mexican, Swedish. With over 30 years R&D expertise, Loquendo is at the forefront of the global speech technologies market. Its high-quality, high-performance technologies (Loquendo TTS text to speech engine, Loquendo Embedded TTS text to speech engine, Loquendo ASR automatic speech recognition) and platforms guarantee the best solutions in 15 languages and 33 voices - with more coming soon. From its headquarters in Turin, Italy, Loquendo deploys technologies that power over 2,000,000 calls every day in the telecommunications and enterprise markets throughout the world. www.loquendo.com Contact: Ornella Ambrois Marketing Manager, +39 011 757 6850 Ornella.Ambrois@LOQUENDO.COM

Loquendo Launches Loquendo Audio Mixer: Loquendo, the global speech technology company, presents the latest development for Loquendo TTS technology: the new audio mixer feature. Using simple command tags embedded in the text, users can now reproduce audio files and music synchronized with the words pronounced by synthetic voices in U.K and U.S English, Spanish, German, Italian, French, Brazilian, Portuguese, Swedish, Mandarin Chinese, Catalan, Mexican, Chilean, Argentinean and Greek. Commands such Mix, Play, Stop, Pause, Resume, Loop and Fade allow users to have complete control on the audio sources. No technical knowledge is required. Thanks to explicit tags among the words, easy synchronization between audio files and speech is guaranteed even if the text is modified. Every sound effect is treated as an independent track, with independent timeline, volume and sample rate. No more offline audio resampling is required: the sample rate frequency of the audio sources is automatically converted according to the voice frequency used. As part of a synthetic speech system, the Loquendo Mixer Audio module allows the realization of extremely high quality, multilingual speech applications. Listen to Loquendo Audio Mixer: Part One Part Two See Mixed files code. www.loquendo.com/ Gruppo Telecom Italia - Direzione e coordinamento di Telecom Italia S.p.A. Contact: Ornella Ambrois Marketing Manager, +39 011 757 6850 Ornella.Ambrois@LOQUENDO.COM

Provox Technologies Corporation announces the appointment of Sheldon Newman as Chief Executive Officer. Shelley Newman was previously CEO of the Misys Physician Systems business unit and prior to that was Senior Vice President of the Misys Enterprise Solutions operations. TuVox recently announced the addition of two industry veterans, Terry Shough and Rick Davison, to lead worldwide sales efforts. Terry Shough joins TuVox as senior vice president of worldwide sales and Rick Davison as vice president of western region sales. Voxware, Inc. (OTC: VOXW), announced that Paul Commons has joined the Company as Vice President and Chief Financial Officer. Sterling Audits announces the appointment of Dianne Waton as CFO. Amir Amane has left Telelogue where he had been CTO and a co-founder.

COMMfusion and Tern Systems are proud to announce the immediate availability of their 2004 Voice/Data Recording market report: The voice/data recording, monitoring, and logging market is one of the fastest growing markets today. End-user revenue is projected to grow to almost $3 billion by 2008 with a CAGR of 13.4% for the time period 2003-2008. COMMfusion and Tern Systems have collaborated to offer "Voice/Data Recording: Markets, Products & Suppliers," the most comprehensive analysis available on the voice/data recording and monitoring industry. This report analyzes the market from 2000 through 2008, providing data and forecasts for the two major market segments – Enterprise Call Center and the Public Environment. The report breaks down these market segments to provide concise, detailed market data from a number of key perspectives: The Enterprise Call Center: The report sizes the market for both full-time call logging and recording, used for Transaction Verification/Confirmation; and call monitoring, used for Quality Assurance and Quality Monitoring. For each subsegment, market data is provided on a variety of metrics, including:

Vendor market share based on systems shipped and CSR shipments in each subsegment are presented through 2003. Data is also provided for revenue per port, revenue per system, revenue per CSR, and ports per system. A breakdown of the Enterprise Call Center value chain is presented for the period 2000 through 2008. In addition, niche/emerging segments within the Enterprise Call Center, such as Call Mining and Third Party Verification are presented and analyzed.

The Public Environment portion of the recording industry is broken down into major segments, including: Lawful Intercept; Air Traffic Control (ATC); Military; Public Safety; Incarceration Monitoring and Public Event Recording. Each of the Public Environment sub-segments is broken down by end-user revenue; system shipments and port shipments. A breakdown of the Public Environment value chain is presented for the period 2000 through 2008.

"Voice/Data Recording: Markets, Products & Suppliers" is the most comprehensive report that is commercially available for voice/data recording in both the call center and public environment market segments. We offer a money-back guarantee that this is so. "Voice/Data Recording: Markets, Products & Suppliers" is available through Tern Systems for just $1,200.00. For additional information, contact Walt Tetschner at TETSCHNER@aol.com or 978-266-1966.

*Supplier*	*Item*	*Q1, 2004*			*Q1, 2003*	*Q4, 2003*
*Supplier*	*Item*	$M	*2002 to 2003 % change*	*Q2, 2002 to Q3, 2003* *% change*	$M	$M
Fonix	Total Revenue	1.93	226.3	173.1	0.590	0.705
	Net Income	(2.34)	45.2	18.3	(4.27)	(2.87)
	Cash or equiv.	2.93	14,555.0	757.0	0.02	0.34
Intervoice	Total Revenue	43.50	13.9	4.1	38.20	41.80
	Net Income	3.20	118.0	(11.1)	(17.80)	3.60
	Cash or equiv.	40.90	56.0	6.0	26.21	38.60
Ivoice	Total Revenue	0.099	(32.8)	5.7	0.147	0.094
	Net Income	(1.55)	(392.8)	(298.9)	(0.32)	(0.39)
	Cash or equiv.	7.06	56.6	245.4	4.51	2.04
Wizzard Software	Total Revenue	0.097	(6.8)	(28.2)	0.104	0.135
	Net Income	(1.28)	266.9	98.8	(0.35)	(0.64)
	Cash or equiv.	0.97	1,767.1	2,157.8	0.05	0.04

*Company*	*Symbol*	*Shares*			*2003*		*Market Cap.*		*Mkt Cap./Rev.*	*2002*
		*Last*	*52-wk.*		*Rev.*	*Net Income*	*7/1/04*	*% change*		*Rev.*
		*Last*	Hi	*Low*	$M		*7/1/04*	*% change*		$M
Andrea Electronics	AND	0.15	0.90	0.13	5.1	(4.6)	7.19 M	(11.78)	1.52	7.2
Apropos Technologies	APRS	3.55	6.61	2.25	20.1	(5.6)	61.30M	(13.21)	3.04	20.9
Artisoft	ASFTC	2.25	4.73	1.36	6.6	(7.1)	8.60M	1.11	1.88	4.5
Aspect Communications	ASPT	13.94	19.95	3.40	363.8	(29.0)	0.815B	3.56	2.48	396.1
Brooktrout Technology	BRKT	10.20	23.18	6.67	74.7	0.6	133.26M	7.14	2.77	73.5
Captaris	CAPA	6.46	7.50	3.85	83.3	12.5	207.53M	3.53	1.78	94.6
Cognitronics	CGN	4.15	4.50	1.85	10.3	(5.4)	23.63M	17.27	1.67	11.3
Comverse Technology	CMVT	19.10	20.87	13.14	765.9	(5.4)	3.74B	11.98	4.59	735.9
Convergys	CVG	14.99	20.80	13.71	2,288.8	171.6	2.15B	2.87	1.05	2,286.2
Covalent Group	CVGR	3.62	4.31	2.10	20.8	(0.6)	47.77M	(4.99)	1.14	29.2
Eckoh (UK)	ECK	10.00	17.70	6.00	104.8	(16.4)	49.4	(9.09)	0.88	104.8
Enghouse(Canada)	ESL	20.98	20.75	12.35	34.6	7.2	197.2	1.34	5.16	13.9
Fonix	FINX	0.263	1.59	0.08	2.4	(13.5)	23.04M	5.21	7.40	3.1
Glenayre	GEMS	2.21	4.30	1.26	58.2	1.6	147.14M	(10.88)	3.65	67.4
Interactive Intelligence	ININ	6.00	6.70	2.67	51.2	(5.9)	95.02M	2.56	1.71	47.8
Intervoice	INTV	11.30	18.00	5.07	156.2	(66.4)	405.98M	(18.29)	2.80	145.6
iVoice	IVOC	0.0009	0.038	0.0004	0.4	(2.0)	3.6	(50.00)	35.33	0.6
NMS Communications	NMSS	7.02	9.35	1.52	87.1	(36.0)	326.42M	(17.89)	2.98	76.5
Nuance Communications	NUAN	4.49	9.40	4.17	55.0	(19.3)	159.4M	(5.27)	6.88	44.1
Preferred Voice	PFVI	0.180	0.33	0.05	2.1	(1.3)	3.31M	(33.40)	0.84	2.2
Prosodie (France)	PRD	21.78	28.21	14.80	203.5	(7.5)	196.7	(4.47)	0.93	203.5
SAFLINK	SFLK	2.59	6.77	1.98	2.0	(10.7)	81.2M	1.25	51.60	2.0
ScanSoft	SSFT	4.80	6.50	3.32	135.4	(5.5)	502.51M	(6.98)	4.79	106.1
TALX	TALX	24.24	31.75	18.50	126.1	13.0	331.36M	3.32	2.25	137.8
VoxWare	VOXW	0.040	0.225	0.035	4.6	(6.6)	20289M	(16.67)	0.93	4.6
Verint	VRNT	32.76	35.89	18.84	192.9	17.9	1.0B	5.04	5.62	131.2
Voiceware (ROK)		0.92	3.67	0.89	3.8	(1.6)	10.5	(17.58)	3.07	3.8
West Teleservices	WSTC	26.25	27.90	20.30	988.3	88.0	1.77B	1.14	2.06	820.2
Witness Systems	WITS	12.33	15.21	4.01	108.04	(20.6)	282.47M	(6.23)	2.79	67.69
Wizzard Software	WIZD	2.70	4.39	0.37	0.470	(2.0)	65.33M	30.06	94.77	0.346
Z-Tel	ZTEL	1.35	4.79	1.25	289.1	(16.1)	65.33M	(0.11)	0.54	235.3
Total	Total				6,630.5	(278.4)	12,931.0	5.08	1.95	5,877.9

The Emperor is Naked!
Comments on Vocalabs paper Say it like a real agent would Ordering Prescription Drugs The Status Quo isn’t going to hack it	Is VoiceXML/SALT to blame? Just what is mainstream? ScanSoft says I’m wrong
Telephone IVR – Applications
Eckoh Launches with VUE Cinemas Cyracom/Envox/Nuance In Healthcare Intervoice Crew Scheduling at CSX National Rail Enquiries Speech service Nuance at Virgin Atlantic	STM Using Nuance in Montreal Syntellect at Camelot for Lottery TeleTimecard service launched Roxio Selects TuVox Pegasus Solutions using Voxify
Telephone IVR – Products/Services
SIS using Acapela TTS Datria Obtains two more patents Intervoice Suite of Applications Launched Vocantas Compliance OnCall Released	SupportSoft/Tell-Eureka Partner TuVox CVR 4.0 Released Vocent DecisionMaker Introduced
Telephone IVR – Partnerships/Financials
IVoice Invests in Corporate Strategies One Voice awarded patent in China Persay/DTSC in VAR agreement	TuVox sales surge Unveil Technologies Raises additional $6 million Viztel IPO raises funding
Telephone Messaging - Applications/Products/ Partnerships/Financial Partnerships/Financial
IP Unity Video Chat Introduced TELUS Deploys IP Unity for messaging Telephone Research for Blind FTW/Kirusa Partner	OnMobile Indian Language ASR ScanSoft Introduces OpenSpeech Attendant ScanSoft Releases SpeechAttendant 8.1 UbiCall at Belgium Ministry of Finance
Telephone Operator Services - Applications/Products/ Partnerships/Financial Partnerships/Financial
Volt Delta Acquires Nortel Networks DA	Phonetic Systems/V-Enable at US Cellular
Telephone Tools - Products/ Partnerships/Financial
Aculab Integrates with InfoTalk ASR Audium Supports Nuance 8.5 ClickFox Customer Monitoring Edify Developer Network Invores Speech-enabled SIP Loquendo/Citec in Agreement	ProNexus Get-up to Speech SandCherry/LumenVox Partner SandCherry Selects Paraxip Media Gateway ScanSoft OpenSpeech DialogModules 2.0 Vocalocity/Excel Switching Partner
Embedded Telematics – Applications/Products/Partnerships/Financial
ART in LG Electronics LX5550 Phone Code Factory using Acapela for Mobile Speak Clarity at Indianapolis 500	Clarity Additional Patent Award VoiceSignal Languages in Samsung Phone VoiceSignal Predicts Voice Text Messaging
Embedded Appliances – Applications/Products/Partnerships/Financial
Spok21 using Acapela TTS for assistive	Loquendo/LARTEC Partner
Desktop – Applications/Products/Partnerships/Financial
Swedish Institute using Acapela TTS IProgrss using Acapela TTS for e-learning InfoVox Desktop v.2 released SYNFACE for Deaf Kurzweil Edu Kurzweil Version 9	MedQuist Using Handheld/PDA Technology ChartLogic Selects ScanSoft DNS Philips at Aiken Regional Medical Center Philips dictation in Hungary ViewPlus using TTS from Wizzard
Data Entry – Applications/Products/Partnerships/Financial
Vocollect Expands Channel Program
Base Technology – Products/Partnerships/Financial
European DIVINES Project Loquendo Releases ASR 6.3	Loquendo Audio Mixer
Industry People
Market Reports
Voice/Data Recording Market Report from Tern Systems/COMMfusion
Financials
Table F1 - Q1 Reported Financial Results	Table F2 – Market Price Summary- 7/01/04
Calendar

Automatic Speech Recognition

*ASRNews***

Pegasus Solutions using Voxify

Kurzweil Edu Kurzweil Version 9