PEABODY, Mass., March 26, 2002 - ScanSoft, Inc. (Nasdaq: SSFT), a leading supplier of imaging, speech and language solutions, today introduced two products based on new AudioMining™ XML speech indexing technology - the AudioMining Development System and Dragon® MediaIndexer™ for Windows. The products use ScanSoft's advanced speech recognition capabilities to automatically create XML speech index and timestamp data for every word spoken within rich media files, and allows text-based keywords and phrases to be used to playback speech information at precise locations within audio and video content.

In a related press release today, ScanSoft announced that Microsoft and a number of leading technology organizations have provided their support for ScanSoft's AudioMining XML speech indexing technology. ScanSoft's new XML speech indexing technologies and products are based in part on audio mining technology previously owned by Lernout & Hauspie (L&H) and recently acquired by ScanSoft subsequent to its December 2001 purchase of speech and language assets of L&H.

"There is a wealth of valuable content that is trapped in the world's audio and visual files," said Jackie Fenn, vice president and research fellow at Gartner Group. "Speech indexing and audio mining is an emerging application of speech recognition that will open up this content to corporate, academic and government customers."

ScanSoft is the first speech and language vendor to enable the affordable cataloging of rich media content, replacing costly manual indexing with automated solutions. The company's new products, one designed for Web and application developers, and the other as a desktop application, are also the first products to enable the integration of XML speech indexing with Web, content management and business productivity applications. ScanSoft's AudioMining XML speech indexing products include:

AudioMining Development System - Allows developers to integrate XML speech indexing, search and precise playback capabilities with Web, commercial, government and custom software applications. The system can be used to add rich media archive and retrieval capabilities to a wide range of applications, including Web, call center, content management, customer relationship management, database, security, training and video editing programs. The AudioMining Development System supports a wide range of applications through its ability to create XML speech index data from broadcast-, telephony- and amateur-quality speech content. For additional information on the AudioMining Development System, please visit www.ScanSoft.com/AudioMining.
Dragon MediaIndexer - A turnkey Microsoft Windows application that creates XML speech index data from tapes and from digital media files in Microsoft Windows Media, RealMedia and other popular formats. The product replaces manual indexing and cataloging efforts with automatic speech indexing and cataloging. The Dragon MediaIndexer is intended to be used by non-technical staff to automatically index and archive media content for Web sites as well as for broadcast, corporate and government archives. For additional information and demo of Dragon MediaIndexer, please visit www.ScanSoft.com/MediaIndexer.

An early customer of the Dragon MediaIndexer is Radio Free Asia, a not-for-profit corporation that broadcasts news and information in nine languages to listeners in Asia. The organization is currently using the Dragon MediaIndexer to automatically index and catalog their archives of news recordings in order to make them searchable through a standard Web interface. The organization had previously been funding an outsourced manual indexing effort, which was expensive, slow and resulted in a limited set of search indexes being keyed into the system. The new approach is affordable, fast and results in 100% of the speech information within rich media files being used as index data.

"Dragon MediaIndexer dramatically increases our effectiveness in distributing streaming news content throughout our organization because it allows our broadcasters and news reporters to instantly find the exact audio or audio visual information they seek," said David Baden, chief technology officer at Radio Free Asia. "Without Dragon MediaIndexer, locating and cataloging our news source material would be a drain on our valuable human and financial resources. This highly effective productivity tool is the ideal solution for our organization, and we have been very impressed with the results."

More About AudioMining XML Speech Indexing

ScanSoft's AudioMining XML speech indexing technology uses an advanced speaker-independent recognition engine to create XML word, timestamp and meta-data for every word spoken within rich media files. XML speech index data makes the speech information within rich media files visible to text-based Web crawlers and search products, unlocking the information hidden within digital audio and video files on the Web and within private media archives.

XML speech indexing eliminates the time and cost associated with manually indexing rich media, enables the indexing of 100% of the speech information within audio and video files, and integrates with standard text-search products to enable rapid access to specific audio and video content.

"The AudioMining applications demonstrate ScanSoft's expertise in combining advanced technology with ease-of-use to allow organizations to overcome barriers to productivity," said Robert Weideman, vice president of global marketing. "These unique and patented products also signal that the former L&H technologies are back in the lead, and that ScanSoft has emerged as a leader and innovator among speech automation vendors."

Pricing and Availability

Dragon MediaIndexer is expected to be available within 60 days for $15,000 per index workstation. The product supports XML speech indexing for U.S. English audio and video content and works with 32-bit Microsoft Windows operating systems.

The AudioMining Development System starts at $25,000 per developer workstation, with runtime licenses required to deploy applications created with the system. The product supports XML speech indexing for U.S. English audio and video content, and works with 32-bit Microsoft Windows operating systems. The AudioMining Development System is expected to be available from ScanSoft within 60 days.

Please visit www.ScanSoft.com/AudioMining for more information and live demonstrations.

About ScanSoft, Inc.

ScanSoft is the leading supplier of imaging, speech and language solutions that are used to automate a wide range of manual processes - saving time, increasing worker productivity and improving customer service. For more information regarding ScanSoft products and technologies, please visit www.ScanSoft.com.

Trademark reference: ScanSoft, AudioMining, Dragon, MediaIndexer and NaturallySpeaking are registered trademarks or trademarks of ScanSoft, Inc., in the United States and other countries. All other product names or company names referenced herein may be to the trademarks of their respective owners.

The statements in this press release that relate to future plans, events or performances are forward-looking statements that involve risks and uncertainties, including risks associated with market trends, competitive factors, and other risks identified in the Company's SEC filings. Actual results, events and performance may differ materially. Readers are cautioned not to put undue reliance on these forward-looking statements that speak only as of the date hereof.

Contact Information