In the span of just a decade, computers have gone from failing to differentiate cats from dogs, to outperforming 92 percent of humans on the SAT. It’s no longer difficult to imagine a world where AI models, trained on vast datasets, can outperform humans in most tasks we consider quintessentially human.
This technological leap enables human-AI partnerships across insurance operations, where carriers combine machine learning models with human expertise to assess and price risk. The advent of generative AI equips carriers with new tools to reimagine countless aspects of their business. This will result in more personalized pricing, enhanced customer service, new strategies to combat fraud, and countless further innovations yet to be imagined.
Building upon this, Sergey Filimonov and the Matrisk (now Akur8 Discover) has spent the last year developing AI tools to analyze insurance filings. Working closely with carriers, we’ve gained insights into the practical challenges of deploying AI within the industry. This essay, published on CAS's (Casualty Actuarial Society) website, shares those findings, along with strategies for successful human-AI partnerships, exploring implications for the insurance sector.
Continue reading on CAS's website.
Full Image Credit goes to Ben Cottier, Robi Rahman, Loredana Fattorini, Nestor Maslej, David Owen
A regular speaker at CAS conferences, Sergey translates AI into actuarial practice for carriers nationwide. His essays on AI have been reshared well beyond insurance—by Google's executives, the founder of Llama-Index, and other industry leaders—highlighting his standing as a cross-sector authority on AI. Before Akur8, he scaled production ML models across Honda’s U.S. operations.
