Blog
All posts
When AI makes producing anything nearly free, the scarce skill becomes deciding whether the output is any good. On why judgment — not technique — is the work.
AI is most dangerous when it's wrong and sure of it. On epistemic risk, and how a knowledge base nobody questions quietly rots from the inside.
Most logistics and distribution businesses lose margin in the paperwork around the move, not the move itself. Where it leaks, what AI can and can't do, and how a rollout works.
Most law and accountancy firms have a leakage problem: unrecorded time, unbilled WIP and documents rebuilt from scratch. Where it leaks, what AI can and can't do, and how a rollout works.
Utility contractors and field services firms have a paperwork problem, not an engineering one. Where time and cash leak, what AI can and can't do, and how a rollout works.
Spend is a governance problem. The models aren't the bottleneck — context is. And companies are done with both decks and tools.
Continuous, AI-driven visibility across a portfolio is finally real. What it does, where it stops, and how to get value from it without drowning in alerts.
Every agent rollout we've run has stalled in the same place. Not model quality. Not context windows. Permissions.
A loop that forgets isn't a loop. It's a repetition. The missing primitive in most AI deployments is somewhere to put what was learned.
When producing the right answer becomes cheap, the bottleneck moves to specifying it. Almost no company has written theirs down.
The car industry agreed on what autonomy means before it built it. Companies are automating without that vocabulary — so here's ours.
Netflix's CPTO says it's the top skill she hires for. We'd go further: most companies can't think in systems because they can't see their own.
Most specialist contractors don't have a labour or materials problem — they have a coordination problem. Where the money leaks, what AI can and can't do, and how a rollout actually works.