A few weeks ago I sat down with the owner of a firm who told me,
flat out, "We don't really use AI here."

Twenty minutes later we were looking at his team's setup together. Three of his people had ChatGPT open. One had pasted a client's full application into it "to summarize it faster." Another was using it to rewrite client emails — names, account details, the works.

He wasn't lying to me. He genuinely didn't know. And that's the whole problem.

Here's the thing nobody tells you about AI in a business like yours: you didn't decide to adopt it. It adopted you. Microsoft switched Copilot on inside the Office you already pay for. Your phone system quietly added an AI notetaker. And your best people — the ones who care, the ones trying to move faster — started pasting client details into Claude and ChatGPT on their own, because it made their day easier.

Nobody held a meeting about it. Nobody signed off. It just… happened. Quietly, helpfully, everywhere.

The thing most likely to put your clients' data
somewhere it shouldn't be isn't a criminal in a hoodie—
it's the tool your most dedicated employee turned
on to do their job.

When I walk a business through this, it almost always comes down to three things they couldn't see:

None of this shows up as a problem on a normal day. That's exactly why it's dangerous. It only becomes a problem when a client asks where their information went, or a carrier asks how you're handling it — or a regulator does.

And on that last one: the regulators haven't written some brand-new AI rulewith a grace period. The obligations you already have — DFS Part 500 ifyou're in insurance, the FTC Safeguards Rule if you're in mortgage —already cover how your business uses AI, right now, no runway. (Worth confirming the current specifics on the regulator's site before you quoteme.) Translation: "we didn't know it was happening" is not the position you want to be defending.

I'm not telling you to ban AI. That ship sailed, and frankly the tools aregenuinely useful. I'm telling you that you can't govern what you can't see —and most owners I meet have never actually looked.

So here's the one thing I'd do this week. Don't send a
policy. Don't write anyone up. Just ask your team,
honestly: "What AI tools are you using to get through
the day?"

Then listen. Don't correct anyone. I promise the list is longer than you think— and the fact that you don't already know it is the real finding.

That's the whole point of this newsletter, by the way. Once a month I'mgoing to show you what I actually see inside agencies — the stuff thatdoesn't make the headlines but quietly ends careers and renewals. No fear-mongering, no pitch. Just what's real, from someone who's in the room.

If the AI question above made you a little uneasy, that's good. It meansyou're paying attention. And if you want to know exactly where your agencystands, that's literally what I do — but that's a conversation for when you'reready, not a button at the bottom of an email.

Talk next month.
— Walter

P.S.‍
I'm doing a live session on this on Thursday, July 16, 2026 · 11:00 AM EST, with Howard Kronberg, the E&Oattorney a lot of you already know. We get into what happens when AI makesthe mistake and whose name ends up on the claim. If that's useful, the detailsare here.