Users Welcome Tools to Build Voice-Enabled Apps

FRAMINGHAM (06/23/2000) - FindGoodHelp Inc. aims to help consumers secure a handyman's or house painter's services through its Web site. But for the Boca Raton, Florida, company, those workers are often hard to track down.

So FindGoodHelp plans to use new voice technology products introduced by IBM Corp. this week to translate consumers' text-based requests into voice messages and forward them to workers in the field - with no human intervention required.

"We don't have to change anything on the back end," said Rajeev Arora, executive vice president at FindGoodHelp. "All we've got to do is change the front end so it now spits out VoiceXML instead of HTML."

The voice-enabled products IBM will ship in late summer and early fall are aimed at providing developers with a common set of tools and middleware to build applications that deliver either text or voice content to a wide range of devices - including PCs and cellular phones.

Key among the offerings is WebSphere Voice Server with ViaVoice speech-recognition technology. Voice Server is the latest extension of IBM's Java-based WebSphere application server suite, which features support for the nascent VoiceXML standard. Pricing will start at $15,000.

"For a while, there was a real boundary between the voice and data world. There needs to be a common development environment and a common clearinghouse of information," said Brian Strachman, an analyst at Newton, Massachusetts-based Cahners In-Stat Group. There are many voice and middleware offerings on the market, he noted, but "it's just easier to implement if it all comes from one vendor."

Developers typically build one front end for an interactive voice response system that accesses a database via the telephone and another front end for a Web browser that accesses that same database, he said.

T. Rowe Price Investment Technologies Inc., an investment management firm in Owings, Maryland, uses one database to house the information its customers can access via the telephone or a Web browser. The two channels also share some logic components, but others are separate, said Scott Quartner, a senior systems consultant at the firm. T. Rowe Price must also write different interfaces for its telephone and Web channels, he explained.

Quartner said that "it would definitely be useful" to have common middleware and tools to speed developer training and make it easier to manage the environment.

T. Rowe Price is currently using IBM's ViaVoice speech-recognition technology.