AI readiness checklist for your organization

Almost every organisation is being asked the same question: what are we doing about AI? Plenty of them respond by buying a tool, running a demo and hoping. A few months later, the tool sits unused, and nobody can say what it was for. The cause is rarely the technology. It is usually that the organisation was not ready for it.

Readiness is not about being large or technical. A twelve-person company can be ready, and a ministry with a big budget can be nowhere near. This guide gives you twelve questions to answer together, a simple scoring method and a clear idea of what to do with the result.

How to use this checklist

1. 01Gather four to six people: someone who owns the outcome, someone who knows the data, someone who will use the result, and someone responsible for risk or compliance.
2. 02Answer each question as a group with Yes, Partly or No. Argue about the answers. The disagreements are the most useful part.
3. 03Score Yes as 2, Partly as 1 and No as 0, and total each section.
4. 04Read the section on what to do next, based on where you scored lowest.

Section 1: data

• Do we know where the relevant data lives, and who owns it?
• Are we legally and contractually allowed to use it for this purpose?
• Is it accurate, current and complete enough that we would act on it?

AI amplifies the quality of what you give it. If your records are scattered across spreadsheets, inboxes and individual laptops, the first project is not AI at all. It is gathering and tidying the data, and that work is worth doing anyway.

Section 2: people

• Is there a named business owner who wants this outcome and will make time for it?
• Have we asked the people who will use the result what they need?
• Do staff have the basic skills to use AI tools and to question what they produce?

Technology projects fail when the people who must change how they work were never part of the conversation. An owner with authority, and users who feel heard, are the best predictors of adoption.

Section 3: process

• Which specific task or decision will change, and how will we measure the improvement?
• What happens when the AI is wrong or unavailable?
• Who will maintain the solution after launch?

A good use case replaces a clearly defined, repeated task. “Improve customer service” is a wish. “Cut the time to answer routine policy questions from ten minutes to two” is a project.

Section 4: risk and governance

• What personal, confidential or sensitive data is involved?
• At which points must a human review the result before it is used?
• Who is accountable for decisions made with the help of AI?

Governance sounds heavy, but at the start it can be one page: what data may be used, which tools are approved, where human review is required and who to ask. Writing it early prevents expensive arguments later.

Reading your score

• 18 to 24: you are ready for a focused pilot. Choose the use case and start.
• 10 to 17: you can start, but fix your lowest-scoring section in parallel, ideally as part of the pilot.
• 0 to 9: begin with a short readiness project: organise your data, name an owner and agree a first use case. Do not build anything yet.

Turning the result into a plan

1. 01Shortlist three to five possible use cases, and score each for value, how easy it would be and how risky it is.
2. 02Choose one with high value, modest difficulty and low risk. Avoid the grandest idea for the first project.
3. 03Write the success measure and the baseline today, for example the current time per task.
4. 04Set a short timeline, often six to ten weeks, and a clear point at which you will decide to continue, change or stop.
5. 05Name the owner, the users and the person who will review the AI’s work.

Mistakes we see often

• Choosing the tool before defining the problem.
• Starting with the most complex process in the organisation.
• Treating the pilot as a demo, with no baseline and no decision at the end.
• Leaving staff out until launch, then wondering why nobody uses it.
• Ignoring who will look after the system after the project team moves on.

What good looks like after three months

A ready organisation does not necessarily have a sophisticated AI system after three months. It has one working pilot, a measured result, a short list of lessons and a decision about what comes next. That is a strong place to be, and a far better one than a large, unfinished programme.

A worked example: a 30-person nonprofit

A nonprofit that runs health programmes wants to try AI. A small group completes the checklist and scores itself.

• Data: 4 out of 6. Programme data is held in one shared system, and the team trusts it, but consent rules for sharing it are unclear.
• People: 5 out of 6. The programme director is eager and has time, and staff are curious, but nobody has used these tools before.
• Process: 3 out of 6. The team can name tasks to improve, but has no baseline time for any of them, and no one owns maintenance.
• Risk: 2 out of 6. There is no AI policy, and no agreement about where staff may paste information.

The total is 14, which suggests starting a pilot while fixing the weakest section. The group writes a one-page AI use policy, ruling out the use of any personal data in public tools. It picks one task, drafting donor reports from field notes, records how long it takes today and sets a six-week pilot with the director as owner and a named reviewer. The weakest section became the first action, and cost nothing but a morning.

Frequently asked questions

Who should be in the readiness workshop?

A decision-maker, someone who knows the data, someone who will use the result and someone responsible for risk or compliance. Four to six people is ideal.

Is a low score a reason not to proceed?

It is a reason to start somewhere different. A low score points to the first thing to fix, which is often cheap, such as naming an owner or writing a policy.

How often should we repeat the checklist?

After each pilot, and at least once a year. Your scores and your needs will change.

Do we need a data scientist?

Not for a first pilot with ready-made tools. You may need one later for custom work, and the pilot will show you what to ask for.

Run the checklist again after the pilot. Your scores will have moved, and so will your understanding of what you really need.