Digital that shows up in the valuation.

I have been working with the machinery underneath this for most of twenty years: parsing and extraction, anomaly detection and pattern analysis, natural-language generation, screening for fraud, long before any of it was called AI. More recently I have run it inside businesses as strategy, governance and systems in production, and I still write code with these tools most weeks. The two things are related. Using them is how I tell a real opportunity from a demo.

What I've settled on, after a year of the same conversation with owners, boards and investors, is a position rather than a strategy. Autonomy should be earned: by an agent from the people who supervise it, by a person through the work that used to be the curriculum, and by a business from the vendors it would otherwise rent its judgement from. A business has to own four things for that to be true: what it knows, what acts in its name, the bar each agent has to clear, and the tests built from its own files.

The whole argument, with the evidence attached and the places it can be proved wrong marked, lives at its own address.