Two Thousand People Trained
Every insurance executive reading this has sat through the meeting. A nice deck on the wall, a chart going up and to the right, and somebody announcing that two thousand people finished the AI training. Everybody claps. Jay Bell has one question for that room, and he asked it when we sat down live at ITC 2026. What can those two thousand people do today that they couldn't do before?
Jay is COO and Head of Business Transformation for Global Construction and Surety at Zurich North America, and he got to that question the long way. Six years as an enlisted Marine. One of the first DSL installers in the country. Then he started over at 29 as a workers' comp claims adjuster at The Hartford, making about a third of what he'd made at the phone company. He traded his Six Sigma skills for an underwriting education and worked his way up to running The Hartford's middle and large commercial programs group. He has done the jobs he's now trying to improve, and it shows.
His point is that the question isn't hard to answer if you ask it before the rollout. Decide how you'll measure it. Calibrate it. Get a baseline. Then train people. Almost nobody does that step, so the chart ends up measuring activity and everybody in the room calls it impact. "I've started to get more relentlessly focused on the difference between activity and impact," he told me.
Same Spot, Same Effort, New Tool
Jay also pointed at where the industry keeps aiming its efficiency work. Intake on the claims side and intake on the submission side, the same two places carriers have poured money into for twenty years. Now leaders want to know where the return on all that AI spend went, and his answer is a fair question back. Why would you expect something different from the same effort pointed at the same spot, when the only new thing is the tooling? He doesn't think a lasting edge comes from your tech stack or your vendor selection at all. It comes from your people. AI is a tool, he said, not the tool.
So Zurich started with people, and with no mandate and no money. A volunteer group called DAILY, short for Driving AI Literacy, passed fourteen hundred members in about three months. One member of his team, Miles Wemken, now Underwriting Operations AI and Middle Market Lead for Construction, was a casual user six or eight months ago and mostly used AI to polish emails. Then the team needed time studies done. Miles spent about sixty hours of his own time over a couple of weeks building an agent in Copilot, and it's now used across Zurich, including in some international markets. It helped that leadership put the tool in everybody's hands. Zurich North America COO Barry Perkins rolled out Copilot Premium to the whole company. What Miles built is a real capability, and you won't find it anywhere on a training completion chart.
One more thing Jay said stuck with me. He loves Six Sigma and says it's at the root of everything he does, but he learned back in 2005 that the harder he leaned on the tools and the jargon in front of underwriters, the more they pulled back. So he keeps the method behind the scenes. Zurich also prefers to staff operations with people who have done the underwriting job, because operations serves underwriting, and it's hard to serve somebody well if you've never walked a mile in their shoes.
The Other Jay Said the Same Thing About Cyber
Two weeks before Jay Bell's episode, we published another conversation recorded live at ITC, this one with Jay Vinda, Global CISO and Cyber Risk Engineering Lead at Mosaic Insurance. Different Jay, different job, and almost the same argument.
When I asked what keeps him up at night, he didn't reach for nation states or some exotic zero-day. He said the basics, for a reason most security buyers would rather not hear. A lot of the controls that actually protect a company are process controls. They don't need a big technology budget, but they take real effort before they'll stop a threat, and the temptation is to skip the effort and buy a tool. His example was data classification. Companies keep trying to hand that problem to somebody else, it gets very expensive, and it still doesn't get solved. Put the time in yourself, he said, and you can solve it pretty quickly. He had a line for the expo floor too. "There's so many tools out here that say they're AI-powered cybersecurity or next-generation AI cybersecurity, and it's just a lot of fluff."
Here's the part that should bother every security buyer. Jay doesn't think frontier AI is a game changer for criminals yet, because they don't need it. Social engineering is cheap and it still works. And the barrier to entry is gone. Cybercrime runs as a marketplace now, with ransomware operators who will handle the business side for you, so you don't need technical skill or a commercial mind anymore. Mostly you need funding. If the attackers are winning with the basics, the defense has to start with the basics too.
He also owns the outcome in a way I wish more executives would. The industry spends a lot of time teaching employees what a bad email looks like, and Jay worries that training has quietly become a liability shift. If you trained them, the click is their fault. He doesn't buy that for a second. "If Mosaic gets taken down because someone clicks on a dodgy link, that's my fault." That matters more every month, because generative AI erased the old tells. In parts of Asia, people used to catch phishing because the Japanese or Korean read like a non-native speaker wrote it. That stopped working almost overnight. The broken grammar and the strange formatting are gone, and often the first message has no link and no urgency at all. It looks like an email meant for somebody else, you reply to correct the sender, and a friendly conversation starts from there.
Why This Sounds Familiar
I've spent twenty-five years building software for insurance companies, and in most of the claims organizations I work with, workers' comp especially, we can solve a lot of the problems without AI at all. I said so on the show with Jay Bell. My favorite book on Lean is 2 Second Lean, which argues that shaving two seconds off every small task adds up to real results. Nobody issues a press release for two seconds, and you only get them by doing the work.
The bootstrapper in me recognizes this too. The whole argument of Be Your Own VC is that people who have had to break even every year, without the safety net of outside money, behave differently. They can't buy their way out of a problem, so they learn to work their way out of it, and they know what every dollar did. The two Jays are describing that same discipline inside big organizations. Zurich didn't buy literacy. Mosaic can't buy a classified data set. Somebody had to do the work.
Cyber is also the chapter in Geek Out, my forthcoming book on seven years of the podcast, where I argue the insurtech thesis actually paid off. The cyber MGAs that came through the hard market were the ones that could prove their monitoring and underwriting discipline produced better loss outcomes, and the mature version of the product looks like a security service with an insurance policy wrapped around it. Jay Vinda's job is close to that idea with a title on it. As Mosaic's CISO he reports to the COO, and as its risk engineer he reports to the head of cyber underwriting, helping about forty cyber underwriters in six countries look at every risk the way an attacker would. He works with insureds on their own security strategy too, which matters more now that Mosaic has opened up to companies in the ten to fifty million dollar revenue range, where there usually isn't a CISO at all. He does for Mosaic's insureds what he does for Mosaic.
What I'd Ask Before the Next Rollout
If you're about to roll out an AI tool, or buy a security product with AI in the name, write down one capability first. Something specific that a person or a control will be able to do in ninety days that it can't do today. Measure where it stands before you start. Then measure it again on the same scale when the ninety days are up. If the number doesn't move, the tool didn't change anything, no matter what the adoption chart says.
If you want a head start, Jay Bell wrote a short field guide on exactly this. It's called The Goat in the Machine, and the goats on the cover, Jellybean and Peanut, are two of the seven he keeps at home. As he put it, "A goat does not care about your strategy." The book includes a nine-question assessment you can run with your team, and he'll give you a free copy at thegoatinthemachine.net as long as you actually run it and send him feedback. That's a guy holding himself to his own standard.
But he asked the question better than I can. What can they do now that they couldn't before? Ask it before the purchase order goes out, and ask it again ninety days later with the baseline sitting on the table in front of you.


