Illustration: Dale Edwin Murray

How would you like to spend the next hour on the phone with an insurance agent?

We can see you recoiling from your screen as if it were made of poison oak.  And yes, everybody here feels the same way too.

“That’s kind of the point, right?” says Anwar Haneef, whose job is to figure out how cognitive computing can do the seemingly impossible: Turn the average person’s experience with insurers into a downright pleasurable interaction.

Haneef is the Watson Global Insurance Industry Leader and Partner for IBM, which means he spends a lot of time helping some of the world’s biggest insurance companies understand how the computer that won Jeopardy! can eliminate the headaches everyday people endure when trying to buy life, auto, home, commercial, and other policies.

Watson, through cognitive computing, can analyze massive amounts of information and pair it with everything from the words we speak to it over the telephone to the comments we type into our social media feeds. Watson then combines all of that information with thousands of other data sources and turns it into something simple and useful—something that, in this case, makes it easier and faster for you to get insurance that’s tailored specifically to your needs.

Getting to that nugget of information is a process that may leave a junior insurance agent stymied, even after three hours on the phone with us, but Watson can complete the task in minutes. In some situations, Watson can eliminate our need to talk to an agent altogether, letting us instead interact with the insurance company through our favorite smartphone app or other device.

“These capabilities can transform the way companies do business,” Haneef says. “You can potentially not have to use the expensive distribution channel of agents anymore.”

Making insurance information available more quickly and cheaply and in a more personalized way is going to be critical to the insurance industry’s survival in the digital age. As things stand today, buying insurance is kind of like the old way of booking a vacation. We would call to make an appointment with a travel agent, have a meeting to discuss options, maybe have a second or third meeting until the travel agent figured out something perfect for the family, and then, finally, book the trip. This could take days, even weeks, so most of us didn’t shop around. Doing so would have meant starting the process all over again with yet another travel agent.

That entire ordeal, which was commonplace not long ago, feels downright archaic in the age of websites such as and, where we hoped to get the best prices and book whatever we want quickly. And yet, the process of buying insurance is still surprisingly similar to the travel agency model—so much so that, in a recent IBM Institute for Business Value survey, 41 percent of respondents said they had left insurers that were too slow to react to their changing needs. It’s a percentage likely to climb as more digital-native millennials begin to seek out insurance products.

Haneef and his IBM colleagues looked at this big picture and found three broad ways that Watson can improve the process. First, Watson can transform engagement, meaning it can, among other things, let us talk to the insurance company whenever, however, and on whatever device we choose. Second, Watson can advise our advisors, meaning it can do all the information-gathering and process it so our time is not wasted during the 10 minutes we actually are willing to spend on the phone with the agent. Third, Watson can help with day-to-day insurance company operations, meaning it can transform something like complex insurance underwriting into a job that even a beginner can do, which would lower the cost of insurance for everyone.

Take that last task, underwriting, as an example. Insurance underwriters receive an application, and then they spend as much as half their time gathering more information. They figure out which information they have, what is missing, and what guidelines to follow regarding the company’s level of acceptable risk.

With Watson doing the information gathering and analysis, the underwriter has a faster, clearer, and more focused picture of each case. What used to take an underwriter with 20 years’ experience a good three days to figure out can now take an entry-level underwriter just a few hours. The insurance company’s costs are reduced, which means the price of insurance should go down at the same time that the insurance product gets better and more personalized for the consumer.

“Over time, if you automate more and more of this, you can transition so that underwriting itself is automated,” Haneef says. “If you can automate 70 percent of the applications coming in, that is going to be hugely profound to the industry. That could change the way organizations go to market. It could change the whole pricing structure in the industry.”

It also frees the most senior-level agents to deal exclusively with larger and more complex clients. For instance, in Singapore, Watson is currently working with a bank to advise wealth managers on how certain events affect their high-net-worth clients’ portfolios. Watson can make connections that the wealth manager might miss among daily global news events, stock-price fluctuations, and other information sources, and then offer the advisor suggestions on how to best help the client.

IBM was able to quantify how much of each manager’s time Watson saved: two hours every day. That’s a full workday’s worth of time being added to every manager’s workweek. At the same time, the managers had more sophisticated information to present to clients, so the clients felt their service was even more personalized.

“It’s always relevant, it’s always up to speed, it’s always up to date,” Haneef says of the Watson system. “And the wealth manager comes across as being somebody who is very knowledgeable.”

The idea of applying Watson to the insurance and banking industries is only about 18 months old within IBM, but the company already has more than 70 clients worldwide, including most of the major banking and insurance brands that consumers would recognize. Most of the companies, though, don’t acknowledge their work with IBM, even if a little marketing might go a long way toward making customers feel better about picking up the phone.

Haneef says the under-the-radar approach is intentional. While everyday people may find the results of a Watson system hopeful, the insurance and banking companies themselves don’t want to give away their tools of competitive advantage.  

“We tested Watson for one of our clients in selling insurance directly online, and what we’ve seen is that this client—which is a leader as it is—have seen a 9-point improvement in conversion with people buying insurance online,” Haneef says. “That might sound small, but this is an industry where a 9-point improvement is something like a 100-percent improvement from what they’ve done in the past, before they optimized. For some of these companies, that translates into hundreds of millions of dollars or even billions of dollars.”