AI Readiness Snapshot

Find the first thing that needs to be true before AI can help.

Ten questions. No email gate. No vendor magic. The result gives you a practical read on workflow fit, data readiness, governance, adoption capacity, and pilot value.

Self-check

Answer from the operating reality, not the desired future state.

This is not a maturity certification. It is a fast way to expose whether the next move should be workflow discovery, data cleanup, governance design, adoption planning, or a bounded pilot.

0 of 10 answered

1. How clearly can you name the workflow where AI should help first? We have pressure to use AI, but no specific workflow yet. We have several ideas, but none are clearly prioritized. We can name one or two workflows with obvious friction. We know the workflow, owner, pain point, and expected outcome.

2. Can the current workflow be explained as steps, handoffs, decisions, and exceptions? Mostly no; it lives in people's heads. Partly, but the exceptions and decision points are fuzzy. Mostly yes, with a few gaps that need interviews. Yes; the workflow is documented well enough to test a change.

3. Where does the information needed for this workflow live? In email, documents, spreadsheets, and memory with no clear source of truth. In several systems that do not agree without manual cleanup. In known systems, but access, quality, or integration work remains. In governed systems of record with known owners and access paths.

4. If an AI workflow used this data tomorrow, how confident would you be in the inputs? Not confident; the team already mistrusts the data. Somewhat confident, but manual verification would be heavy. Confident for a limited pilot with clear checks. Confident; quality rules, ownership, and review paths are understood.

5. Do people know what information can and cannot go into AI tools? No shared rule exists. People use judgment, but rules are informal. We have written guidance, but it needs workflow-level detail. We have clear rules for public, internal, confidential, and regulated data.

6. Who is accountable for reviewing AI-assisted output before it affects a real decision? Nobody has been named. It depends on the person using the tool. A likely reviewer exists, but the evidence trail is not designed. Reviewers, authority boundaries, and evidence expectations are explicit.

7. How much real attention can the organization give a first AI pilot over the next 90 days? Almost none; everyone is already over capacity. A few hours here and there, if nothing urgent interrupts. A named owner can spend time weekly with support from stakeholders. A named team can meet, test, decide, and remove blockers quickly.

8. How has the team handled recent process or technology changes? Changes tend to stall, fade, or create workarounds. Adoption works only with repeated follow-up and manual enforcement. The team can adopt changes when the reason and support are clear. The team regularly improves workflows and expects measured change.

9. Can success for the first AI use case be measured without hand-waving? No; the expected value is still vague. We can describe value, but not yet measure it. We have a practical measure such as cycle time, rework, throughput, or quality. We have baseline measures and a decision threshold for continuing or stopping.

10. If the pilot worked, would it connect to a real operating decision or business outcome? Not yet; it would mostly be a demo. It would be useful, but ownership after the demo is unclear. It supports a real outcome, with some production-path questions left. It supports a real outcome and has a plausible production owner.

Find the first thing that needs to be true before AI can help.

Answer from the operating reality, not the desired future state.

Readiness tier

Copied summary

Pillar scores

Three practical next moves