The AI readiness audit: what I actually check

Every AI consultancy engagement should start with a readiness audit, and most of what gets called that is really a tools demo dressed up as an assessment. The version that's actually useful checks the business's foundation, not its enthusiasm.

The six things I check: data accessibility, meaning can the relevant data actually be pulled via API or does it live in someone's head, process documentation, existing tool sprawl, a realistic ROI baseline, and leadership sponsorship beyond the person who initiated the project.

Most audits fail on data accessibility or process documentation first: not because the business lacks the data or the process, but because nobody's ever needed it in a structured, accessible form before, and that gap is invisible until someone specifically checks for it.

The audit's real value isn't the report. It's that it surfaces the actual blocker before six figures get spent discovering it the hard way, three months into a project that was never going to work without fixing the foundation first.