Trust follows leadership trust
A comment last week gave me a harder question than any post has this month: how do you make recognition of someone's new value structurally likely, not just possible?
It came from a reader pushing back on "own the new capital": if the organisation doesn't build room where new capital can matter, re-skilling is a trick, not an invitation. Fair. Then came the follow-up — can you build a system for that?
I don't think you can, and organisation design shows where it stops. Phanish Puranam's framework treats rewarding contributions as one of four problems every structure must solve — pay, promotion, praise, all designable. But the moment someone says, in front of others, "your call changed this" — no design table produces that.
What you can build is narrower: put people with the new capital where decisions get made, and make someone accountable for naming, out loud, whose judgment actually moved the outcome. Recognition stays a human act. Structure only decides whether anyone with authority is in the room to perform it.
That's the exact shape of this chapter's problem. 66 percent of people use AI regularly; only 46 percent are willing to trust it — a twenty-point gap, KPMG and the University of Melbourne, 2025, 48,340 people across 47 countries. The instinct is to close it with a better model.
David De Cremer's research says otherwise: trust in an AI system follows trust in the people who introduced it — reliability alone doesn't earn it. Same limit as recognition — you can't systematise your way into it. Somebody has to carry it, visibly, on purpose.
One move: stop tracking "AI adoption" as a usage number. Ask who, this quarter, is accountable for naming out loud when a person's judgment overrode the system — and was right. Put a name on that, not a metric.
The tool can remove the excuse of not seeing what someone carried. It still can't do the carrying, or the naming. That's a person's job, three chapters running now.
Next Tuesday: what it looks like when two halves stop taking turns and start owing each other.
💬 Who in your organisation currently holds that job — naming when a human's judgment beat the system? Or does no one?