FCAT Scans for Best Technologies

The entrance to the fidelity center for applied Technology in Boston is emblematic of the groups mission to seek out emerging technologies: Every month or so, a new biometric device guards the door. In early February, a facial recognition device stood watch.

“We try a different one every month—face recognition, iris scan, thumbprint, hand geometry,” said Charlie Brenner, senior vice president of FCAT. “We have not yet had the Aha! experience.”

But thats OK. Brenners job is to check the stuff out and see if its got potential. If it passes muster with Brenner and his team, he hands it to others to build production versions.

“We try to stand between our business units on the one hand and [vendors] labs on the other and identify the things that make sense,” said Brenner, whose unit tests for ease of use, security, compliance with standards and—above all in these economic times—value.

There are several technologies in which Brenner perceives value at the moment. One of them is grid computing. He said he sees the emerging technology not as pie in the sky but as eminently practical, a way to enhance reliability—always important in financial services—through redundancy. He also said he sees it as a way to utilize the thousands of servers that Fidelity has purchased.

“Its a cloud,” said Brenner, pointing out that, in that sense, grid computing is like the Internet itself, which can route around failures. “Were looking to see if we can take those concepts and apply them to transaction processing, load balancing and availability,” said Brenner.

Another pursuit with potential—although its a few years from realization—is called Natural Broker. In a nutshell, its a brokerage workstation that could process spoken input from online customers and render it on a visual display.

“People are better off speaking and reading,” as opposed to writing and listening, Brenner said. “No one has ever combined those things in production.”

Although FCAT developers produced a working prototype in 2001, the cost of building it in volume for the mass market would have priced it beyond the reach of customers, Brenner said. Its time will come when Moores Law brings about microprocessors powerful enough to perform speech recognition in the handheld device itself, he said.

One area where FCAT research has paid off already is in audio mining technology. In one instance, digitally recorded conversations with customers are converted to text and then scanned for keywords using a natural language search tool. “Its helpful if we want to know if customers are asking about the new lifetime savings plan,” said Brenner.

Fund managers can also use the tool to scan quarterly earnings conference calls for keywords, said Brenner.

Brenner will continue to search for, well, a better search tool to guide customers through Fidelitys Web site. To many customers who know what they want and want it right away, conventional menu-based navigation is a barrier.

“So you need a search technology that is so robust that it becomes the interface of choice,” said Brenner. In that, a natural language tool—one that can understand the way people speak every day—is key.

Its all in line with carrying out Brenners mission of seeking deep understanding of user needs—and auditioning new technologies to see which is ready to serve them.