Insights
Notes on making AI work in practice
Short pieces on what decides whether AI gets used: choosing the right problem, fitting the work, and the operating model around it.
-
Natural-language queries are only the interface
Asking data questions in plain English is not enough. The bigger opportunity is moving analytics from pull-based reporting to push-based decision support, and that depends on context.
-
Many AI projects get evaluated the wrong way
Five questions I use to estimate whether an AI project will actually succeed: trust, workflow fit, adoption, safe scaling, and defensibility.
-
Enterprise AI's happy ears problem
A strong demo or an interested sponsor does not mean a use case is real. The hardest part of enterprise AI is usually problem selection.
-
We have seen this movie before: dashboard hell, now with AI
Analytics teams learned that tools fail when they skip decisions, users and context. AI projects are repeating the same mistake, with higher stakes.
-
AI is repeating the business intelligence story
Twenty years ago, companies wrestled with the same questions about BI that they now face with AI. The lessons carry over: a strong central backbone with ownership inside the business.
-
AI succeeds or fails on the operating model, not the model
Five practices that separate organizations that scale AI from the ones that stall: platform thinking, clear ownership, security in charge, and building for reuse.
Start with a 30-minute conversation
Tell me where AI stands in your company today. I will tell you plainly where I think it can help, and whether I am the right person to help with it.
Email me to set up a 30-minute call