Chase Manhattan Corp. has adopted a voice-recognition system for its bearer bond processing department.
Chase is the biggest bearer bond processor in the country and sees voice input as a big time-saver, said Nicholas Papinakolaw, Chase's senior vice president and chief operating officer for worldwide technology and operations, global trust.
Since the bank went live with the Ume Voice Inc. software last year, processing time per coupon has been reduced from 17 screens in seven minutes to one screen in less than one minute, and training time for employees has dropped from one month to less than three days.
Eighteen people have been trained to use the system.
The bank took the Ume Voice technology and integrated it with its own record-keeping system so that it has a "purely Chase Manhattan focus," said Mr. Papinakolaw.
Now Chase is planning to roll out the system from its securities center in New York City and data processing center in Dallas to sites in New Jersey and in London and Bournemouth, England.
"We are in the process of reviewing additional applications," said Mr. Papinakolaw. "In fact, wherever we have installed a keyboard we could in theory replace it with a voice application," he added.
He declining to say how much the bank has invested in the technology. "That's not an issue," he said. "It has absolutely paid for itself."
Ume Voice, based in Novato, Calif., was founded by Adithya M.R. Padala, its president and chief executive officer, more than two years ago. That was six years after it had developed voice-recognition technology for UmeCorp., its parent, which started out as an artificial intelligence company.
Ume Voice is selling voice-driven software to banks for back-office processing, custody and settlement, and cash management; and to brokerage and dealer houses for order entry, execution reporting, currency, equity and bond trading in the front office.
"We knew volumes of transactions would increase so substantial performance improvement was necessary," Mr. Padala said.
Ume Voice executives said they would like to set the voice-recognition standard for Wall Street.
Their product is designed to recognize different intonations and accents but can also be "trained" to work for only one operator.
"Apple first came out and created graphical user interfaces. Now we're creating voice user interfaces," Mr. Padala said.
"Our niche is the desktop application. Instead of typing, a person talks to the system in the desktop environment."
The system can recognize numbers and stock symbols, draw graphs and tables, perform analytics, and connect in real time to securities markets and news wires.
Along with Chase Manhattan, the company said it is working with two top brokerages.
"Banking users are different from trading users," Mr. Padala said. "Traders are high-energy, highly educated, highly paid, and impatient. On the bank side, users are generally high school graduates who may have language barriers and could be scared of computers."
Judith Markowitz, a voice-recognition and verification consultant based in Chicago, said Ume Voice "has a very good track record." Users "are positive about Ume Voice and have a good relationship" with the company, she said.
Mr. Padala's goal is to have the majority of the top 10 financial institutions signed by next year. Internationally, the company may install the system, in Japanese, at Nomura Securities in Japan.
"We've created a buzz in London," said Paul Cavise, Ume Voice's vice president of marketing and business development. "With all the mergers going on, we're going to focus on our clients in the United States and then go global with them and establish ourselves overseas."
