Thinking
What I'm working through
Essays on AI, organisations and brand: how trust and attention are won, moved and destroyed. Each one first appeared on LinkedIn, where the discussion lives.
Enterprise AI map
Twenty-four agents across ten functions, the trade-off each one creates for everyone else, and what decides whether faster work turns into a better organisation.
AI transformation: faster work, different constraints?
Once a team can do in an afternoon what used to take a week, what happens to the work afterwards matters rather more. Does someone use it? Does it wait for a decision? The companion piece to the map: the work got faster, the constraints moved.
We keep firing the honest fool
A model that tells you what you want to hear is not a bug in the technology, it is a preference we trained into it, and one organisations have been training into people for a lot longer. Amsterdam, 1637: the man who called the tulip market right got run out of the room.
At a buy-and-build, creating a strong brand is often a missed opportunity
There is a confident genre of content right now: the fractional CMO who unlocks EBITDA for private equity. Almost every piece frames brand the same way, and almost every piece misses where the value in a buy-and-build actually sits.
The coordination problem no agentic roadmap has priced
Part three. Every company can rationally decide not to train the juniors it no longer needs, and the industry still ends up short of seniors in a decade. Nobody's roadmap has a line for the cost of that.
Juniors don't build judgement by overseeing AI
The practical sequel: watching a model work no longer forms anyone. Run a gap analysis over the competency grid you already have, then build a deliberate second loop that routes real calls back to juniors so judgement still gets made. Aviation learned this the hard way and put manual flying back on the schedule.
If AI takes the junior work, where does the next senior come from?
AI is eating the execution layer from the bottom, the rung where people used to start. Keep juniors on the payroll and you can still produce a generation that never earns judgement. A system can inherit judgement; it cannot make more of it.
The CEO AI panic of 2026 is a governance failure, not a leadership one
Most CEOs now fear their job rests on AI paying off, and boards are filing the shortfall under leadership. The real issue sits one seat over: boards underwrote a cost-cutting bet, then grade the CEO on a growth one. That is a governance failure, not a leadership one.
Marketing is where the operating model breaks
A response to HBR's agentic-marketing piece. Marketing is not the bottleneck because marketers are slow; it is where a broken operating model first becomes visible. The orchestration layer everyone calls efficiency is really a question of who owns the work, who carries the risk, and whose budget it sits in.
The hidden politics of AI adoption
The meeting was supposed to be about choosing an AI platform. Twenty minutes in it had become something else: who signs off, whose headcount shrinks, whose budget it comes out of. Most adoption does not stall on technology.
The shortlist effect has a dark side nobody's discussing
Thomas Marzano’s Shortlist Effect: if an AI agent doesn’t shortlist your brand, you don’t exist. I largely agree. The part nobody discusses is lock-in; once agents learn your preferences, how does a new brand ever reach you?