There are those who advise, and those who execute. The AI consulting market long lived off that split. I think it's now costing companies dearly.
The slide disconnected from the field
An AI strategy written by someone who has never deployed a model is a map drawn by someone who has never walked the terrain. The recommendations are often right on paper, and unworkable in reality.
What coding changes
- I assess real feasibility, not theoretical.
- I immediately see the integration and data pitfalls.
- I can deliver a prototype right after the scoping.
- I speak as an equal with your engineering teams.
What I recommend, I can build. And what I can build, I don't oversell.
The dual profile as a safeguard
When the strategist and the developer are the same person, there's no loss between intent and deliverable. No "it should have worked". The recommendation is tested by the person who made it.
That, I believe, is the most honest form of AI consulting: the one that exposes itself to the reality of production.
