In short
An AI readiness checklist is a short set of pass/fail tests that shows whether a business has the data, process clarity, ownership, rules and people AI needs before money is spent. The eight signs below are symptoms a leader can spot in daily operations, each with a test that can be run within a week.
Key takeaways
- Readiness is judged one use case at a time, and this checklist is a first screen.
- Each of the eight signs has a pass/fail test using real records and real people.
- Fix the named use case and its owner first, then scope the rest to it.
- Gartner, RSM and Cisco surveys show readiness gaps are common, though each has limits.
Most businesses fail a readiness review on ordinary things: data nobody trusts, processes nobody has written down and no one accountable for the result. A structured AI Opportunity & Readiness Assessment examines these in depth. This post is the self-check version, part of the AI implementation playbook.
What is AI readiness, and how common are the gaps?
Business AI readiness means a business can name a specific problem, trust the data behind it, run the process consistently and name someone accountable for the result. An AI readiness framework or AI maturity model rates the whole organization. A checklist asks whether one use case can start safely.
The gaps are common. Gartner asked 1,203 data management leaders in July 2024, and 63% said their organization lacked the right data management practices for AI or was unsure. RSM's 2025 survey of 966 middle market leaders found 91% already used generative AI, yet 61% of US firms said they were not fully prepared.
Cisco's 2025 AI Readiness Index, covering more than 8,000 leaders in 30 markets, put about 13% of organizations in its top Pacesetter group. Cisco sells networking equipment, so its findings carry a commercial interest.
What do signs 1 to 4 look like in data and process?
Data and process gaps show up as arguments about numbers, repeated manual fixes and steps only one person understands. A model fed by that work inherits the same confusion. These four signs cover data readiness for AI and the process underneath it.
| Sign | Pass/fail test |
|---|---|
| 1. Nobody agrees where the real numbers live | Two people produce one monthly metric separately. Pass if the figures match. |
| 2. Data quality is assumed, never measured | Count gaps in 50 records the use case would read. More than 5 fails, a threshold this checklist proposes. |
| 3. The process lives in people's heads | Three people write the steps and exceptions. Pass if the main steps agree. |
| 4. AI is wanted in general, with no named problem | One sentence names the task, owner, today's cost and the measure. No number, no pass. |
RSM found 41% of users with rollout problems named data quality first; Cisco found only 58% of organizations had a defined AI roadmap.
What do signs 5 to 8 look like in people, ownership and budget?
People and ownership gaps show up as silence: no named owner, no rules staff can quote, no word to the team whose work will change, and a budget that stops at the license. These four signs cover organizational AI readiness, which technical reviews often skip.
| Sign | Pass/fail test |
|---|---|
| 5. No one is accountable for the outcome | Ask three senior people who answers for a wrong output. Pass if they name the same person. |
| 6. Staff use AI tools, and no rules exist | Ask five employees what they may enter into AI tools. Pass if answers match a written rule. |
| 7. The affected team has not been told | Ask five team members what would change in their week. Pass if four can answer. |
| 8. The budget stops at the license | Pass if budget lines exist for data preparation, training and ongoing ownership. |
Cisco found only 35% of organizations had a change-management plan, and 39% of unprepared firms in the RSM survey cited missing in-house expertise.
How do I score this AI readiness checklist and choose what to fix first?
Run all eight tests on one candidate process and count the fails. The count is a working guide, not a validated score, but on how to know if your business is ready for AI, named fails beat instinct. Fix signs 4 and 5 first: the use case and its owner set the scope of the rest.
As a rough reading, 0 to 2 fails suggests scoping a pilot while fixing the rest, 3 to 5 suggests fixing the failed signs first, and 6 to 8 suggests a wider review: an AI maturity assessment, digital readiness assessment, AI capability assessment or staff AI readiness survey.
Passing every test does not guarantee a good pilot; the post on why AI projects fail covers what else goes wrong. Where Sarvadhi fits. When several tests fail, Sarvadhi's readiness assessment gives an outside reading of the gaps and prioritized opportunities.
Conclusion
An AI readiness checklist turns a vague worry into eight tests with a clear pass or fail. Most failures are ordinary: scattered data, unwritten processes, no owner, no rules and a budget that stops at the license. Fixing them before a pilot costs less.
Run the tests on one process this week and put a name beside every fail. If several fail, or an independent view would help, Get an AI Readiness Assessment.
Filed under AI Strategy Consulting