Insuranceciooutlook

Artificial Intelligence: Empowering People - Not Machines

By Michael Freilich, EVP & Chief Technology Officer, Renaissance Alliance

Michael Freilich, EVP & Chief Technology Officer, Renaissance Alliance

The last time I wrote about artificial intelligence (AI) in the insurance industry was in 2017. I discussed the spectacular possibilities of intelligent automated interactions, personalized service for complex issues, and new distribution channels. Looking back on the article, I feel some of the same optimism I expressed two years ago—but I also see my naivetéat the time.

In some ways, the promise of machine learning appears even more spectacular today. We’ve seen remarkable achievements in game-playing in systems like AlphaZero and AlphaStar, realistic pictures of people “imagined” by generative adversarial networks and the image processing power of driverless cars and trucks. These remarkable technologies inspire a wealth of ideas for greater efficiency and effectiveness in the workplace.

In the insurance industry, in particular, AI applications are becoming increasingly common; it is rare to find an insurance startup that does not use the words “machine learning” in its pitch deck. However, I am seldom impressed by these uses of technology. The reality is that AI is readily available to any startup. Google’s decision to open-source TensorFlow—arguably one the most transformative decisions of the software age—made the power of machine learning available to anyone with some knowledge of computer programming. It is not surprising that new businesses are capitalizing on the availability of this technology and the excitement that surrounds it. 

Perhaps even more importantly, the ideas I hear for AI applications in insurance rarely take into account the current state of affairs on the ground. In my work at Renaissance Alliance supporting independent agencies with their technology needs, I see rampant use of antiquated software that is difficult to use and requires multiple clicks to complete simple tasks. I observe a widespread failure to leverage the power of basic programs, a prevalent perception of a need for outdated tools, and a general lack of the time and capital necessary to invest in more sophisticated technology and to learn to use it effectively. At a more fundamental level, many agents are struggling to measure their business operations and translate those measurements into actions that will help them grow. The majority of the AI solutions that are currently on the market do not address these challenges.

"In the insurance industry, in particular, AI applications are becoming increasingly common; it is rare to find an insurance startup that does not use the words “machine learning” in its pitch deck"

And while we are fortunate that entrepreneurs and investors have set their sights on insurance, pouring capital, creative ideas, and talented individuals into the business, it is clear that AI is not yet ready to replace people. Computers can simulate empathy in some narrow situations, but they cannot yet truly connect with people. A wise colleague occasionally reminds me that insurance is a sad business; people pay premiums, and they get nothing in return until something horrible happens. Success in this sad business requires building real human connections that are grounded in empathy. We’re not close to creating an empathetic machine that can interpret subtle clues about human emotion and respond with compassion. These computers cannot, therefore, replace a human voice, particularly during the times of confusion or crisis in which compassion and patience are vital. In fact, several startups in this space that were initially focused on providing AI chatbots have come to appreciate this problem and have pivoted to “platform” plays. They found that the real value of the tool they built was in how it enabled content delivery and connected people—not in the chatbot itself.

AI-driven automation then is a red herring. The insurance business today needs people—not AI—to deliver critical messages. And we need technology that can empower people rather than replace them. Rosalind Picard, the founder, and director of the Affective Computing research group at the MIT Media Lab, articulates this need and the true potential of the technology: “We’ve been doing a lot of reflecting to figure out how to build tools that will enable a better future for humans, not for machines. When we build technology that reminds people that they matter, that helps them achieve something they couldn’t do before, I just feel joy from head to toe.” While Picard isn’t talking about insurance specifically, her philosophy encapsulates how AI should and will transform our industry.

Indeed, there is a place for AI in the insurance agent’s office today, and there can be value in the AI that startups are deploying. But the best use of this technology currently is not for automating interactions—it is for empowering and connecting people. And the real challenge facing the insurance industry is not in producing AI-generated insights, but in delivering these findings to agents in a way that takes into account the real technology limitations and time restrictions they face. The industry is starved for software that is intuitive, responsive, effortless, and fun to use. We need technology that delivers information quickly and seamlessly to the people who use it, that makes the value of analytics easy for agents to understand, and that empowers them to apply its findings to their advantage. Truly, the way in which the software delivers these insights is as important as the data itself.

There is certainly work to be done to make this technology more accessible, but successful use of AI in the insurance industry will require more learning by people than by machines. Insurance agents will need to be open to the possibilities AI creates. And even the most user-friendly technology will require agents to take the time to develop new skills and change their processes to incorporate it. Change is never easy, even when it brings with it the promise of growth.

I have little doubt that one day, AI will transform the insurance industry. In some areas, this transformation has already begun. But let’s leave “intelligent automated interactions” in the laboratory for now. Our focus in the near-term should be on using AI to empower people rather than machines.

Read Also

How to Make a Notoriously Reactive Industry Proactive

How to Make a Notoriously Reactive Industry Proactive

Mike Gulla, Senior Director of Underwriting, Hippo Insurance
Harnessing Power of Blockchain to Restore Trust

Harnessing Power of Blockchain to Restore Trust

Dorota Zimnoch, Founder and MD, ZING Business Consulting

Weekly Brief

New Editions