Read the Room
We expect others to respond to information they don't have. Turns out, so does AI. The problem is older than both of us.
Short essays on enterprise AI, forward deployed product work, customer reality, systems thinking, and the conditions that make execution either easier or unnecessarily painful.
We expect others to respond to information they don't have. Turns out, so does AI. The problem is older than both of us.
The Luddites weren't wrong. They were just living through the part of the story where the ending wasn't visible yet. So are we.
The most interesting question about AI isn't what it can do. It's what it gives us back.
The blockers are never where you expect them. I've known this for years. I just didn't expect my first week at a new job to prove it quite so efficiently.
The biggest blockers to enterprise AI adoption usually aren’t model quality or infrastructure. They’re alignment, trust, workflow design, and organizational ambiguity.