Jonathan Roomer numbers / people / plot twists
  1. Jonathan Roomer is a Chartered Fellow of the Institute and Faculty of Actuaries and co-founder of YuLife, where he built and scaled the customer organisation across three continents and now focuses on applying AI to real-world business problems. Before YuLife, he earned an MBA at Imperial College and led KPMG's Tech Growth team, advising hundreds of startups on how to grow. He speaks on AI, insurance, and the shifts that matter before they become obvious.

I keep getting pulled into what happens when meet and neither behaves as .

  1. I trained as an actuary, so I'll still love and build the spreadsheet. I've just learned not to leave it alone in a room with the future.
  2. People read the rules, spot the incentive, and quietly become something the model never intended. Reward steps and the fitness tracker ends up on the dog. We log that as an error. It's the most honest data in the set.
  3. So I stopped asking what people will do and started asking what everyone's quietly assuming they won't. That second question is where the fun begins.

I go looking for the a makes sense on paper and complete nonsense to a .

  1. Insurance began as the loveliest idea we ever had: we're all in this together, and if it goes wrong for you, the rest of us cover it. Then we got good at it. Underwriting, mortality tables, pricing each person for their own risk. Every step sensible, every step fairer than the last. Follow it far enough and you arrive somewhere strange: a pool of one, where I pay only for me. Perfectly reasonable, and also no longer insurance. We didn't break it. We optimised it until the "we" quietly left.
  2. Take dynamic pricing the same way. Obviously we should reprice people as their lives change. Now that we finally can, why wouldn't we? Then you picture the actual text. "We noticed you spent ninety minutes in an ice cream shop. Premiums up 1.5% until you visit a gym." Nobody designed that to be absurd. Everyone just followed a good idea one ping too far.
  3. My favourite version is prevention. The dream is insurance that stops the bad thing instead of paying for it: fewer heart attacks, fewer claims, everyone better off. But how do you prove the value? Do you send this letter at year's end? "Congratulations, you didn't have a stroke, you didn't miss work, we paid nothing. Premiums up 25%." The better it works, the harder it is to prove it did anything at all.
  4. Which is the bit I keep circling. Every one of these is a good decision followed one step past its sense. I'm not sure any of them is wrong. I just can't stop noticing how a thing built so we'd never face the worst day alone keeps optimising its way back to: you're on your own.

Some of this won't fit in a footnote.

A few things currently refusing to leave me alone:

  • The claim that never happened - this will be my ITC Vegas keynote.
  • The future of insurance was published in 1927 (Huebner). We still haven't built it.
  • If we can't trust anything we see, hear or read, does business go back to being done face-to-face?

If those are your kind of questions, say hello.

  1. Jonathan Roomer is a Chartered Fellow of the Institute and Faculty of Actuaries and co-founder of YuLife, where he built and scaled the customer organisation across three continents and now focuses on applying AI to real-world business problems. Before YuLife, he earned an MBA at Imperial College and led KPMG's Tech Growth team, advising hundreds of startups on how to grow. He speaks on AI, insurance, and the shifts that matter before they become obvious.