IBM's Watson : Today Jeopardy, Tomorrow Insurance Applications

In a brilliant bit of public relations hearkening back to the chess matches between IBM's Deep Blue computer and Gary Kasparov in 1996 and 97, IBM supplied its Watson question answering computer to play against all time Jeopardy champions Ken Jennings and Brad Rutter. In last night's opening round of three, the Single Jeopardy round ended with what IBM blogger Steve Hamm called a "cliff hanger": Watson $5000; Brad Rutter $5000; Ken Jennings $2000.

Hamm relates that at times the human contestants were able to beat Watson to the buzzer, showing the edge that human intelligence still has in some respects over artificial intelligence. Watson showed his sophistication by attempting a correct French pronunciation of Jean Valjean, the protagonist of Victor Hugo's Les Miserable, Hamm reports.

Watson's participation on Jeopardy is being broadcast on the local ABC networks. In the NY area, for example. It is being broadcast on ABC Channel 7 at 7:00 P.M.

What's brilliant about IBM's stunt is that it showcased artificial intelligence capabilities relevant to many industries, including the insurance industry. Everybody who has called 411 or interacted with an automated customer service representative (CSR) will be acutely aware of the limits of such capabilities. Jamie Bisker, IBM's global non-life insurance segment manager asserts that Watson represents a leap forward, both in its natural language text interpretation and the potential of linking those capabilities with sophisticated voice recognition.

"Combine emotional recognition and knowledge synthesis with a Watson question answering system, and you're talking about a competitive advantage," Bisker says.

However, even more straightforward uses of Watson-like technology could add value to insurance operations. For example, Bisker says that claims adjusters could instantaneously to things such as answer a range of questions about company claims history and average payouts involving a certain model of car. Examples could be multiplied in just about any area of insurance operations that could benefit from a tool that can query a body of knowledge in practical ways.

"Insurers could ask themselves, 'What kind of problems do we need answers to," Bisker explains. "Suppose their biggest problem was training new agents, and not being able to always have a senior agent assist a junior one. This kind of capability will give them that extra bit of confidence and be able to supply meaningful answers to questions."

Such tools are yet to be created, but it won't be long before they are, Bisker predicts. He says that possibly within three years, certainly within five, vertical solutions will be sold and implemented. He acknowledged that IBM also expects competitive offerings to emerge in the not-too-distant future.

For a little information on the hardware and data sources that supply Watson's knowledge base, among other interesting information, check out the embedded video below.

Anthony O'Donnell has covered technology in the insurance industry since 2000, when he joined the editorial staff of Insurance & Technology. As an editor and reporter for I&T and the InformationWeek Financial Services of TechWeb he has written on all areas of information ... View Full Bio