Self-assessment
AI readiness assessment: find your next step
For engineers and risk teams: the scoring model and framework mapping
The assessment: 13 questions
Pick the answer closest to where you are today. Your immediate result includes three priorities and a plan you can keep. Answers stay in your browser.
0 of 13 answered
Ownership and value
One senior person owns the outcome, and the value is written down as a number.
The work today
The process AI would change is mapped as it really runs, and it is measured.
Data and systems
The data exists in usable shape, access to it is controlled, and the systems involved can be reached.
Choosing what to build
Ideas are compared on value, feasibility and risk, and buying has been weighed against building.
Controls and people
Staff use of AI is known and governed, sign-off is agreed early, live systems are watched, and users are trained.
Your result
0 of 13 questions answered. Your score and action plan appear here once every question is answered. If this page is not scoring for you, add up your answers (0, 1 or 2 each) and apply the score bands and ownership and control gates below.
What this assessment checks
An AI readiness assessment checks whether an organisation has what it needs before it spends money building with AI: someone senior who owns the outcome, a clear idea of the value, data and systems that can support the work, and the controls to run it safely. The gaps it finds are cheapest to fix before anything is built.
This one scores 13 questions across five areas, following the discovery method of our AI Opportunity & Feasibility Sprint and checked against two public sources: the NIST AI Risk Management Framework, and the UK government's consultation on AI Management Essentials. Your score is worked out in your browser, and copying or downloading the plan sends nothing to us. If you ask for the plan by email, we receive your address and answers to prepare it.
- Ownership and value
- One senior person owns the outcome, and the value is written down as a number. 2 questions.
- The work today
- The process AI would change is mapped as it really runs, and it is measured. 2 questions.
- Data and systems
- The data exists in usable shape, access to it is controlled, and the systems involved can be reached. 3 questions.
- Choosing what to build
- Ideas are compared on value, feasibility and risk, and buying has been weighed against building. 2 questions.
- Controls and people
- Staff use of AI is known and governed, sign-off is agreed early, live systems are watched, and users are trained. 4 questions.
What your score means
- Each question has three answers, scored 0, 1 and 2 in the order shown. The most you can score is 26.
- The initial score bands are 0 to 9: Not ready to build yet; 10 to 18: Ready to plan; 19 to 26: Ready to validate a pilot. These are our planning thresholds.
- Ownership and controls can lower the initial band. Question 1 scored 0 gives Not ready to build yet at any total. A score below 2 on ownership (question 1), access (6), sign-off (11) or monitoring (12) caps a high total at Ready to plan. Missing controls must be established; partial controls need validation before a connected pilot.
- Every answer below 2 is a gap, with a different action for missing and partial preparation. The result shows three priorities first, then the remaining gaps. Ownership comes first; at planning or pilot scores, control gaps follow. At low scores, value and workflow come next. Other gaps are ordered by score, then question number.
| Dimension | Score | What it means | Sensible next step |
|---|---|---|---|
| Not ready to build yet | 0 to 9 | Your answers flag foundational gaps. Establish the sponsor, the value and the workflow before committing to a pilot. | Bring a senior sponsor and the people who do the work together. Choose one process, measure the starting point and locate its data. Use your priorities below as the agenda. A facilitated planning session helps if the decision stalls. |
| Ready to plan | 10 to 18 | Your answers suggest useful foundations, with gaps to resolve before a controlled pilot. Use the priorities below to agree what needs evidence and who will provide it. | Run a short, fixed-scope discovery that ranks your use cases against your real data and systems and ends in a costed plan, judged by people who can weigh the business case and the engineering together. Our AI Opportunity & Feasibility Sprint is that discovery: fixed scope, typically two to four weeks. |
| Ready to validate a pilot | 19 to 26 | Your answers suggest a foundation for validating one bounded pilot. Confirm the supporting evidence and address any flagged gaps before committing to it. | Agree a limited scope, test data, success measures, human review and a way to pause. Validate against the baseline. A production rollout needs separate testing and approval from your accountable business, legal, security and risk owners. |
If your band is Ready to plan, the AI Opportunity & Feasibility Sprint turns your priorities into a costed plan against your real data and systems.
Before you choose, set our services against the four checks to run on any supplier.
For engineers and risk teams
For engineers and risk teams: the scoring model, framework mapping and limits
The model is deliberately unweighted: 13 questions, three ordinal answers each, summed to a total out of 26. Weighting would imply a precision a self-reported score does not have. Ownership and control gates are applied after the total: no senior owner caps the result at the lowest band; partial ownership or access, sign-off or monitoring below 2 caps a high total at planning. These gates prompt further preparation and evidence checks; they do not verify that controls work.
Most questions map to a step of the Sprint's discovery method and, where one fits, to subcategories of the NIST AI RMF 1.0 core. The mapping is our reading of the framework. It is not a NIST conformity check, and NIST does not endorse this assessment. The systems-access question has no direct RMF equivalent; it is kept because the Sprint's feasibility step establishes which systems would have to be touched.
Three questions are not from the Sprint. Question 4 asks for a baseline, because the value in question 2 cannot be checked without one. Questions 12 and 13 come from two gaps respondents named in the UK AI Management Essentials consultation. Those gaps were monitoring AI systems after deployment, and building AI skills across the organisation.
Respondents also noted that organisations using AI as a service often cannot see a vendor's internal processes. If you buy rather than build, ask the vendor what they will disclose.
Limits: answers are self-reported and unchecked, and nothing here tests your data or systems. The bands are cut points we chose, not a benchmark against other organisations.
| Dimension | Where it comes from | NIST AI RMF 1.0 |
|---|---|---|
| 1. Is one senior person accountable for what AI should achieve here, with the authority to release budget? | Leadership sessions | GOVERN 2.1, GOVERN 2.3 |
| 2. Can you say, in money or hours, what the first AI project should change? | Leadership sessions | MAP 1.4, MAP 3.1 |
| 3. Is the work you want AI to help with written down as it actually happens, including the spreadsheets and manual steps? | Workflow analysis | MAP 1.1 |
| 4. Do you measure that work today: how long it takes, what it costs, or how often it goes wrong? | Added: needed to check the value in question 2 | MAP 3.1 |
| 5. Do you know where the data for that work lives, and what state it is in? | Data and systems feasibility | MAP 2.3 |
| 6. Is it clear who may see that data, and could an AI tool be held to the same rules? | Data and systems feasibility | MAP 1.6 |
| 7. Can the systems involved be connected to, and does someone own that connection? | Data and systems feasibility | No direct equivalent |
| 8. Do you have a shortlist of possible AI uses, compared on value, how hard each is to deliver, and risk? | Opportunity map | MAP 3.1, MAP 3.2 |
| 9. For the top idea, have you checked whether a product you could buy already does it? | Build versus buy | GOVERN 6.1, MAP 4.1 |
| 10. Do you know which AI tools your staff already use, and is there a written rule on what data may go into them? | Governance review | GOVERN 1.4, GOVERN 1.6 |
| 11. Is it agreed what your legal, security and risk teams need to see before an AI system goes live? | Governance review | GOVERN 1.1, MAP 1.5 |
| 12. Once an AI system is live, is it agreed who checks its work and what happens when it gets something wrong? | Added: AIME consultation gap | MANAGE 4.1, MEASURE 3.1 |
| 13. Have the people who will use AI been trained on what it does well, where it goes wrong, and the rules for using it? | Added: AIME consultation gap | GOVERN 2.2, MAP 3.4 |
Frequently asked questions
What is an AI readiness assessment?
An AI readiness assessment is a structured check of whether an organisation can start using AI productively: whether someone owns the outcome, the value is defined, the data and systems can support the work, and the controls to run it are agreed. It is done before money is committed to building, so the gaps are found while they are cheap to fix.
What is the difference between AI readiness and AI maturity?
Readiness asks whether you can start. Maturity asks how well you already run AI across the organisation. This assessment is a readiness check: it is most useful before the first or second AI project, and questions 10 to 13 cover the controls a more mature organisation would already have in place.
Is there a UK government-backed AI readiness assessment?
Yes. Innovate UK's BridgeAI programme offers the AI Adoption Assessment Toolkit, built by Digital Catapult for UK startups, scaleups and small and medium-sized businesses. It covers digital maturity, AI readiness, data readiness, data ethics and MLOps maturity. Separately, the Department for Science, Innovation and Technology consulted on a self-assessment tool called AI Management Essentials from 6 November 2024 to 29 January 2025. In its response, published on GOV.UK on 6 February 2026, it said it will not publish that tool, and plans guidance on foundational AI governance aimed at small and medium-sized businesses.
How much does a paid AI readiness assessment cost?
It depends on the scope. One published example is Gartner's G-Cloud 14 pricing document, published in 2024, which lists an AI Capability Maturity Assessment at £142,400 as a fixed-price service. This self-assessment is free. A review of your actual data and systems is what our AI Opportunity & Feasibility Sprint adds: fixed scope, typically two to four weeks.
Do we need outside help to become ready for AI?
Your own team can close the early gaps: naming an owner, putting a number on the value, mapping the current process and listing the AI tools staff already use. Outside help earns its fee where the business case and the engineering have to be judged together: ranking use cases against real data and systems, and designing the controls. That is the work our AI Opportunity & Feasibility Sprint does.
Is this assessment an audit or a certification?
No. It is a self-assessment: the score is only as accurate as your answers, and nobody checks them. It does not test your data or systems, and it is not a conformity assessment against the NIST AI Risk Management Framework or any standard.
What happens to my answers?
Answers stay in browser memory unless you choose optional email delivery, when available. Copy and download happen on your device; refreshing clears the answers on this page. An email request sends your address and answers to prepare your plan through Resend. Our business inbox receives your address and separate updates choice, without your answers. There is no automatic marketing subscription. Our privacy notice explains retention and deletion.
Sources
- [1]1AYM, AI strategy and roadmap: the AI Opportunity & Feasibility Sprint and its discovery method
- [2]NIST, Artificial Intelligence Risk Management Framework (AI RMF 1.0), NIST AI 100-1, January 2023 (read 29 September 2026)
- [3]Department for Science, Innovation and Technology, AI Management Essentials tool: consultation and outcome, GOV.UK, updated 6 February 2026 (read 29 September 2026)
- [4]Department for Science, Innovation and Technology, Government response to the consultation on AI Management Essentials (read 29 September 2026)
- [5]Innovate UK Business Connect, AI Adoption Assessment Toolkit from Digital Catapult, Innovate UK BridgeAI programme (read 29 September 2026)
- [6]Digital Catapult, AI Adoption Toolkit (read 29 September 2026)
- [7]Gartner, G-Cloud 14 pricing document, 2024 (read 29 September 2026)
In this topic
Is AI worth it for us, and are we ready for it?
- AI ROI: estimate it, then check itBuyer guide
How do you estimate the return on an AI project before you build it, and check it after launch?
How much does AI implementation cost, what drives the number, and how should we budget for it?
Where does AI create value in a private equity portfolio company, and what should the first 100 days deliver?
Further
- Data platform & AI enablement · Where to look if your gaps are in the data and systems questions.
- Agent proposes, verifier gates · The control pattern behind the monitoring and sign-off questions.
Bring us your action plan.
In a 30-minute call we go through your priorities with you and agree the next step, whether that is an internal planning session, a discovery or a controlled pilot.
Last reviewed · 1AYM