Last week we discussed which unlikely occupations can make the claim of having more mandated training than independent insurance agents. This week, we're looking at artificial intelligence and the role it will be playing in business in only a few years—and what impact it might have on independent agents and brokers.

As if you don't have enough to worry about with direct writers and online insurance sales sites, now comes IBM's “question answering systems,” as personified by Watson on the ”Jeopardy!” game show.

In case you haven't heard, Watson beat “Jeopardy!” champ Ken Jennings and turned in a generally impressive debut performance as the face of artificial intelligence—or “question answering systems,” as IBM prefers to call it.

Listen to a podcast, “How can IBM's Watson help insurance agents?”

First, a few facts about Watson:

Watson is the result of the work of 25 IBM research scientists over the past 4 years.
Watson was named after IBM founder Thomas J. Watson.
Watson evaluates the equivalent of roughly 200 million pages of content (about one million books), written in natural human language, to find correct responses to complex questions.
Watson was specifically downloaded with information typical of what's used in “Jeopardy!” using Wikipedia and IMDb among its sources.

IBM was pretty smart in rolling out its Watson technology in such a benign way. By debuting on a game show, the technology seems simultaneously non-threatening and so advanced that practice usage seems remote.

But it isn't. In fact, Watson-type analytics technology will be available to the healthcare industry in 18 to 24 months through a recent joint venture with IBM and Nuance Healthcare. Using input from Columbia University Medical Center and the University of Maryland School of Medicine, IBM and Nuance built a database that will allow medical professionals to ask questions in real language and get a quick and accurate answer.

According to Jamie Bisker, industry solutions representative for IBM, this level of technology will be available to the general business community in as soon as 3 to 5 years.

So what does this mean for the future of the insurance agent?

Not the apocalyptic assumption of power by Watson that the more lurid-minded might imagine. In fact, the advance of question-answering technology will be less HAL 2000 and more “Star Trek,” with Watson-like personal assistants serving as your own personal Ken Jennings, Bisker says:

Watson-like systems would be most helpful as assistants or used in the way that research librarians are utilized to help answer questions. An agent may ask a potential policyholder whether they had all their questions answered or not prior to the call, with the assumption being that they used online knowledge-based tools in their insurance shopping much like transactions today.

After all, between 60 and 80 insurance carriers are already using automated underwriting, Bisker said; Watson-like technology will be more like “Google on steroids” than an omniscient, malevolent machine bent on ruining your life.

“Insurance business type questions, such as those around the coverage that a particular product provides or its limitations, could be queried in everyday language by an agent or her staff. This would help keep costs down, and free up human experts to assist in the really tough questions,” he said.

Question-answering systems that were “trained” (provided with the right combination of information) on product language would provide consistent and more accurate answers whenever they were used.

Another advantage is that these system would provide an audit trail of exactly what was asked and the answers that were provided, Bisker said. For tasks like marketing and generating sales plans, an agent could ask the Watson-like system that was fed information about his book of business a series of questions instead of trying to produce a report using a screen. Even the relatively straightforward use of reporting systems in use today would be made easier once a person could simply ask for what they wanted.

Carriers adopting question-answering technology can position its use as another high-tech tool to attract both independent agents and consumers, Bisker maintains.

“An AI-based assistant that was used to answer consumer and agent questions would be more consistent day-in and day-out, would not be susceptible to performance anxiety, the pace of business, or the state of commissions,” he said. “It wouldn't accept bribes (there's an extra kilowatt it in for you if you tell them they need more coverage…) or be too tired to answer a question at any hour of the day or night.”

Another advantage is the lack of “emotional coloring” that Watson brings to the process of insurance placement. Not that the current system lends itself to unethical behavior, but having an impartial Watson to help place coverage a few years back would have saved some big brokers a lot of grief.

“Part of a knowledge-based system could be a set of corporate morals, ethics, values and other statements of conduct that would be used as a final or even initial check of actions, questions or answers for,” Bisker said.

Watson could also come in handy for training new producers. “A carrier or independent agency could leverage such systems to bring new producers online faster and with more confidence if they knew that the 'training agent' was an AI-based tool that could be accessed from the trainee's phone or computer,” he said.

Bottom line, Watson's use of natural language and high level of abstraction makes it a perfect agent helper of the future.

At least that's how the Star Trek school of thought pictures it. For the HAL 2000 camp, the prognostication could be a little grimmer and goes something like this:

Big insurance companies decide to do away with their carbon-based distribution force — why pay commissions? — and place Watsons in their regional offices, downloaded with state- and line-of-business specific data. Easily updated and programmed, Watsons can easily become a customer's best friend, becoming the “trusted advisor” they need in today's complex insurance marketplace.

Which scenario do you think is more likely?

Special Reports /

Toops Scoops: Agents, Are You Smarter Than a Supercomputer?

Related Stories

Recommended For You