"Are we ready for AI?" is usually the wrong first question. The better question is: what happens today when someone in your company needs a number, moves a piece of information, or makes a decision? AI readiness lives in those answers, not in a maturity framework.
After working inside several small and mid-sized operations, the same handful of signals keep predicting whether an AI or automation project will land. None of them mention AI at all.
The five signals that actually predict success
1. Do your numbers agree?
Ask two teams for the same figure, say monthly volume by region. If the answers differ, and reconciling them takes a meeting, any AI tool pointed at that data will confidently produce answers people argue about. Trust in the output can never exceed trust in the input.
2. How long does a new question take?
Not a standard report, a new question. "Which customers slowed down this quarter?" If the honest answer is days, because someone has to pull exports and assemble a spreadsheet, the constraint is retrieval, not intelligence. Fixing retrieval is boring, unglamorous work, and it is exactly the foundation AI needs.
3. How much moves by hand?
Count the copy-paste. Information that travels between systems inside someone's clipboard is information that arrives late, occasionally wrong, and invisible to any tool you buy. Every manual bridge is both a cost today and a blocker for automation tomorrow.
4. Does anyone own the workflow?
Tools get owners. Workflows rarely do. When a process crosses three departments, and nobody is accountable for how it runs end to end, an AI rollout has no one to redesign the process around what the technology can do. Ownership gaps are the quietest project killer on this list.
5. Do improvements get measured?
If your last three process improvements were declared successful rather than measured, an AI project will meet the same fate. A baseline taken before the build, and a follow-up in your own operational data, is what separates a real result from a good demo.
What readiness is not
Readiness is not a score out of one hundred, a vendor questionnaire, or a licence count. It is not something you need to finish before talking to anyone. Most importantly, low readiness is not a reason to wait. The gaps themselves are the work: connect the data, remove the manual bridges, assign ownership, set the baseline. Do that around one workflow and you have both a readiness improvement and a business result to show for it.
How to check yours in two minutes
We built a short self-serve version of exactly these questions: the AI Readiness Self-Check. Nine questions, instant result, and a practical next step for whichever area comes out weakest.
If your answers point at the foundations, that is not bad news. It means the highest-value project available to you is well understood, and it is the kind of work described under Diagnose and prioritize. Bring one workflow and it becomes very concrete, very quickly.