The sound of your voice can be the key that unlocks new IVR systems
Periphonics Corp. of Bohemia, N.Y., which supplied the IRS' TeleFile interactive voice response technology, is turning to voice verification on its IVR systems. Periphonics has incorporated voice verification products from Veritel Corp. of Chicago and T-Netix Inc. of Englewood, Colo., into its Open Signal Computing and Analysis Resource, or OSCAR, speech-processing algorithms. A caller's voice can determine system access rights similar to the way a fingerprint does.
Periphonics Corp. of Bohemia, N.Y., which supplied the IRS TeleFile interactive
voice response technology, is turning to voice verification on its IVR systems.
Periphonics has incorporated voice verification products from Veritel Corp. of Chicago
and T-Netix Inc. of Englewood, Colo., into its Open Signal Computing and Analysis
Resource, or OSCAR, speech-processing algorithms. A callers voice can determine
system access rights similar to the way a fingerprint does.
The financial industry is likely to try voice verification first, and the government
likely will be in the second wave of early adopters, marketing coordinator Jodi L. Keller
said.
Agencies constantly have to do more with less money. They are going to go with
technology that lets them do this, Keller said.
IVR systems lead a caller through an automated menu by prompting responses on a phone
keypad. IVR can route calls and retrieve information from databases and, if
speech-recognition software is installed, it also can respond to spoken words.
This spring the IRS began a nationwide rollout of a TeleFile addition that will let
employers file Form 941 quarterly tax returns.
TeleFile presently verifies taxpayers via their Social Security numbers and separate
personal identification codes. Veritels AuthentiGate or T-Netixs SpeakEZ would
secure access by voice verification alone.
A caller would record a password for storage in the Periphonics Voice Template Database
Server. In later calls, the person would be prompted to speak the password. The signal
would go to an OSCAR running AuthentiGate or SpeakEZ, which would match it against the
stored password.
Tolerance can be set depending on the degree of security desired. At zero tolerance,
even a head cold could invalidate the match. At the other end of the scale, almost any
voice speaking the password would be admitted.
NEXT STORY: Faxes are low on Army's priorities list