Email newsletter
A biweekly field report on enterprise AI deployment, written for IT leaders by the people doing the work.
One real AI rollout lesson from the field
One AI story worth your attention
One practical framework your team can use today
Latest editions
Issue 05
Your AI Didn't Get the Memo
You cleaned up the docs, set up the integrations, and things worked. Then a policy changed — and the AI kept answering from before the update. The model was fine. The pipeline wasn't. Here's what's actually happening, and a two-minute test to find out if it's happening to you.
Read more →
Issue 04
Most AI Tools Are Not Actually Smart
Every AI product markets itself as smart. But retrieval and critical thinking are different things — and most enterprise tools only do one. Including, in some ways, ours. Gartner now has a word for the gap: agent-washing.
Read more →
Issue 03
Personalization Without Transparency is Just Surveillance
I asked Claude to challenge an idea. It pushed back in a way that felt unusually pointed. Then I found out why: it had built a detailed profile of me without telling me. The enterprise version of this problem is worse.
Read more →
Issue 02
In One Document Apples are Vegetables. In Another, Fruits.
The real cause of most enterprise AI hallucinations isn't the model — it's conflicting documentation. When two policy documents give different answers, the AI picks one and sounds confident. Plus: how EY published an AI-hallucinated citation.
Read more →
Issue 01
The Part of AI Deployment That Doesn't Make the Keynote
Most AI deployment writing skips access controls — the part that determines whether everything else works. Who decides what AI gets to see? How does per-user access actually work with Jira? The view from the build side.
Read more →