Hey,

Can I still wish you a happy New Year? I’ll do it anyway!

I’m back. Took December off. Full month. No laptop. No Slack. Just family, books, and long walks.

It gave me time I needed. Time with my family. But also time to think.

Distance does that. When you step away from the daily noise, patterns become clearer. You see things you missed when you were in the middle of it.

And I kept thinking about how AI is showing up in teams. How we’re adopting it. How we’re not designing it.

That pattern became impossible to ignore.

Last week I grabbed a drink with an old colleague. We haven’t worked together in years, but we stayed in touch.

He told me about their new AI agent. Running in Slack. Helping the team. Everyone loves it.

Then he said:

Really smart guy. Good intentions. But no design. And that confirmed what I’d been thinking about all month.

AI feels reliable. Speaks convincingly. Responds quickly. Has an interface that radiates trust.

So we trust it!

Not because we’re naive. But because we’re inexperienced.

We don’t have a reference frame for distrusting AI yet. With software, we know where the boundaries are. With colleagues, we know what can go wrong.

But with agents? No instinct. No experience. No red flags.

This isn’t blind trust. This is first-generation trust. And that makes it dangerous. Innovation without design is improvisation. And improvisation never scales.

During that month off, I kept coming back to one question:

What would a practical design model look like? Something teams could actually use. Not governance theatre. Not compliance paperwork. But real design.

FRAME is my answer.

It’s a one-page design model to determine how AI is allowed to participate in your team and systems. Not what AI can do, but what AI may do.

It’s Five domains. Five questions. On one page:

F – Function (Why does this agent exist?)
R – Risk (What can go wrong?)
A – Agreements (How does it behave in the team?)
M – Mandate (What can it do autonomously?)
E – Evaluation (How do you stay in control?)

I’ve written everything out on a landing page. The full explanation. With a complete template. Everything you need to start using FRAME today.

Pick one AI agent in your team..

Now answer this in one sentence: What is it allowed to do without a human?
If you can’t answer that clearly, you don’t have an AI strategy. You have hope.

Then do the simplest next step: Pick one task that agent should own end-to-end. Add a human approval step where it matters.

What did you just give your agent permission to do?

Write it down.

Let’s make 2026 Human First. Supported by AI. Designed on purpose.

Have a good week!
Jos