Free checklist
The AI Readiness Checklist
The questions to ask before you build, and the failure each one is trying to prevent.
Most AI projects do not fail because the model was not clever enough. They fail because nobody could say what “working” meant, the data was worse than anyone admitted, or the thing shipped with no way to tell when it was wrong.
This is the checklist I work through on a readiness audit, 32 questions across 6 sections, each with the specific failure it is trying to prevent. It is written to be useful whether or not you ever speak to me.
Read it now
Opens straight away, and I will email you a copy to come back to.
What is inside
1. The problem
Everything downstream depends on this, and it is where most projects are already lost. If you cannot fill this section in, no amount of engineering rescues it.
5 checks
2. The data
Nearly every project discovers its data is worse than believed. The only variable is whether it discovers that in week two or month five.
6 checks
3. What good looks like
If you cannot measure it, you cannot tell improvement from noise, and you cannot tell a regression from a bad day.
5 checks
4. Humans in the loop
Nothing leaves the pass without the head chef looking at it. Money moving, emails sending, files writing, code deploying: none of it should happen without an explicit human yes.
6 checks
5. Security and trust boundaries
The moment your system reads anything from outside (a web page, an email, a supplier's PDF) it is reading text that may be trying to give it instructions.
5 checks
6. Running it
The build is the short part. This section is what separates a demo that impressed everyone in March from a system still earning its keep in November.
5 checks
If you work through it and want a second opinion on what you found, an AI readiness audit is the formal version of this: same questions, against your actual systems, with a written answer at the end. Or just tell me which one you got stuck on.