Essays on AI, business, and the work it takes to bring the two together. Published as the practice generates them, usually weekly.
Baker asked something I haven't been able to put down. In your AI setup, which model has the mandate to tell you you're wrong? You have no board. No partner paid to flag when the plan doesn't hold. So you build that seat yourself. And you stay at the top of the stairs.
A head of marketing sent me a deck a few days ago. Slides, copy, layout, narrative arc. Designed and written end to end by Claude. He hadn't opened a single slide before forwarding it. This is the second wall: what breaks once AI is in the building and nobody is checking.
A lawyer without a line of code beat 500 professional developers at a hackathon. She built a complete application faster than people with twenty years of experience. This is what happens when English becomes a programming language.
Three AI agents running on schedule can hand you your entire business day by 7am. Not as noise. As signal. Decisions ready to make, problems ready to solve. This is how fractional operators actually scale.
It's 2006. A first-year analyst at a top-tier London consulting firm, doing work nobody had quite named yet. Twenty years later, that work has a name and AI just changed everything about who can do it.
A CEO asked me which platform her exec team should standardize on. I told her she was asking the wrong question. The right question is which tool earns which work, and how to make that division explicit.
For years, "our data can't leave the building" ended more enterprise AI conversations than any other sentence. Three frontier labs just made that excuse much harder to defend.
Last week, UK financial regulators made something unmistakably clear: AI is no longer an innovation topic. It is a governance topic. That shift matters far beyond financial services. If you lead a business, it should get your attention.
Most executives have tried ChatGPT. Most were underwhelmed. The problem is rarely the model. It is how they use it, and what changes when they stop treating AI like a vending machine.