Answers
What should an AI readiness assessment cover?
Four things, in this order: which use cases are actually worth pursuing given your real constraints, whether your data can genuinely support them today, a clear build, buy, or wait recommendation for each one, and a phased roadmap with success metrics defined before any code gets written.
Use-case selection comes first because it's where most AI initiatives actually fail — not in the engineering, but in picking a use case that sounded good in a meeting and doesn't survive contact with real constraints, real data, or a real cost-benefit case. A good assessment ranks candidates by impact versus feasibility, not by which one is most exciting to talk about.
Data readiness gets checked next, honestly. It's common for a use case to be genuinely valuable and genuinely blocked by data that's incomplete, inconsistent across systems, or simply not accessible yet — finding that out during an assessment costs a few weeks. Finding it out mid-build costs a lot more.
The build, buy, or wait recommendation is the part most assessments skip, and it's often the most valuable line in the report: sometimes the honest answer is that a use case isn't worth custom development yet, either because an existing tool already solves it or because the data isn't ready. A roadmap with real success metrics, sequenced by what's actually achievable first, is what turns all of that into something a team can act on.