AI readiness assessment
Twelve questions, about five minutes, scored the moment you finish. It runs entirely in your browser: nothing is submitted, nothing is stored, and no email address is asked for at any point. The result is designed to be useful even when it tells you not to spend money yet, which for roughly half of the organisations we speak to is the right answer.
Twelve questions
Answer as things actually are rather than as you would like them to be. A flattering score is worth nothing to you.
What this actually measures
Readiness for AI has very little to do with how advanced the technology is and almost everything to do with six unglamorous properties of your organisation. Two questions each, six points available per dimension.
Problem clarity
Whether you are pointing at a specific, measurable problem or at a general ambition.
Data condition
Whether the information a model would learn from exists, is reachable and is trustworthy.
Systems and integration
Whether a result could actually reach the place where the work happens.
Ownership after handover
Whether anyone internally will run the thing once it is built.
Oversight and risk
Whether you know what happens when the system is wrong, and who answers for it.
Sponsorship and funding
Whether the money and the patience exist to get past the first disappointing result.
Notice what is not measured
There is no question about which model you should use, whether you have a cloud strategy, or how much budget you hold. Those matter far less than most vendor assessments imply. What decides outcomes is whether the problem is specific, whether the data exists, whether the answer can reach the work, whether somebody owns it afterwards, whether errors get caught, and whether the sponsor will survive a disappointing first version.
This assessment exists because a large share of the enquiries we receive describe a project that should not be built yet. Not a bad idea, just an idea arriving before the groundwork that would let it succeed. Saying that on a call is awkward. Letting you find it out privately, in five minutes, at no cost, seemed more useful.
The AI readiness assessment is a free twelve question self check covering problem clarity, data condition, systems and integration, ownership after handover, oversight and risk, and sponsorship. Each question scores zero to three, giving a maximum of 36. The result places you in one of four bands, shows which of the six dimensions scored lowest, and gives specific next steps for the two weakest. It runs entirely in your browser, nothing is submitted or stored, and no email address is required.
How the scoring works
Each question offers four answers worth zero, one, two and three points. Twelve questions give a maximum of 36. The bands are deliberately harsh at the bottom, because the most expensive mistake in this field is starting a build that the organisation is not equipped to finish.
| Score | Band | What it means in practice |
|---|---|---|
| 0 to 11 | Foundations first | A build now would probably waste money. Usually the data does not exist in usable form, or nobody owns the result. |
| 12 to 19 | Narrow the scope | Something real is here, but smaller than you are imagining. Cut hard and prove value on one task. |
| 20 to 27 | Ready for a pilot | A properly scoped pilot is a reasonable spend. Discipline now matters more than capability. |
| 28 to 36 | Ready to build properly | You have what most organisations lack. The risk shifts to choosing the right problem. |
Why these questions and not others
Most readiness assessments published by vendors ask about the things that vendor sells. A cloud provider asks about your cloud maturity. A platform vendor asks whether you have a platform. The questions here were chosen from a less flattering source: the reasons projects we have seen actually stalled.
- Data condition comes first because it usually decides the outcome. Across almost every engagement, the gains came from fixing inputs rather than swapping architectures.
- Problem clarity is the cheapest thing to fix and the most often skipped. If you cannot describe how you would grade the system, it is not ready to be built, however good the demonstration looked.
- Integration is the most underestimated line in any budget. A prediction that never reaches the place the work happens changes nothing at all.
- Ownership decides whether it survives. Systems without a named owner degrade quietly and get switched off within a year, usually without anyone deciding to switch them off.
- Oversight is not paperwork. If there is no way to tell a correct answer from a confident wrong one, the system is not ready regardless of every other score.
- Sponsorship decides what happens after the first disappointment. Version one underperforms. Organisations that expected that iterate. Organisations that did not, cancel.
What happens to your answers
Nothing. There is no submit button that sends anything anywhere. The scoring runs in JavaScript in your own browser, the result is written into the page you are already looking at, and closing the tab destroys it. No answer, no score and no identifier reaches us.
Two things are worth naming so the claim is complete. This page carries Google Analytics, which is switched off until you accept it in the banner at the bottom of the screen, and which is the only part of this site that measures anything. If you accept, it records that someone completed the assessment and which of the four bands they landed in. It never receives your individual answers, your score breakdown, or anything that identifies you. Decline and none of that happens, and the assessment works exactly the same. The detail is in our cookie policy.
What to do with the result
If you scored in the bottom two bands, the honest next step is not to talk to us about a build. It is to fix whichever dimension scored lowest, and the result panel tells you which two those are with a specific action for each. Most of that work does not require an external supplier at all.
If you scored in the top two bands, a conversation is probably worth thirty minutes. Bring the result with you, because the weakest dimension is usually the most useful thing to talk about first. Our pricing page explains how engagements are staged so that you can stop after any of them, and the first conversation costs nothing either way.
Frequently asked questions
Marked up with FAQPage schema so these answers can surface directly in search results and inside AI assistant responses.
What is an AI readiness assessment?
It is a structured self check of whether an organisation is equipped to get value from an AI project, as distinct from whether the technology could work. This one asks twelve questions across six dimensions: problem clarity, data condition, systems and integration, ownership after handover, oversight and risk, and sponsorship. It scores zero to 36 and places you in one of four bands, from foundations first through to ready to build properly.
Is this assessment really free, and do I need to give an email address?
Yes it is free, and no email address is required at any point. There is no gated report and no form standing between you and the result. The scoring runs in your own browser, the result appears on the page immediately, and no answer or score is sent to iSpark.
How long does the AI readiness assessment take?
About five minutes. There are twelve questions with four options each, all on one page, so you can scan ahead and go back. Answer them as things actually are rather than as you would like them to be, because a flattering score is worth nothing to you.
What score means we are ready to build an AI system?
Twenty or above out of 36 puts you in the band where a properly scoped pilot is a reasonable spend, and twenty eight or above suggests you have what most organisations lack. Below twelve the honest advice is to fix foundations before commissioning anything. The result panel names the two weakest dimensions and gives a specific action for each.
What are the six dimensions of AI readiness?
Problem clarity, whether you are pointing at a specific measurable task. Data condition, whether usable history exists and is reachable. Systems and integration, whether an answer could reach the place the work happens. Ownership, whether anyone will run it after handover. Oversight and risk, whether errors get caught and who answers for them. Sponsorship, whether the funding and patience exist to survive a disappointing first version.
Are my answers stored or sent to iSpark?
No. The scoring is JavaScript running in your browser, the result is written into the page you are already viewing, and closing the tab destroys it. This page does carry Google Analytics, which stays switched off until you accept it in the banner. If you accept, it records only that an assessment was completed and which of the four bands resulted. It never receives your individual answers or anything identifying you.
Where to go after the result
Whichever band you land in, one of these three is the useful next click.
Data engineering and AI readiness
The dimension that scores lowest most often, and the work that has to happen before a model is worth building.
Read more →Pricing
The three engagement shapes, what actually drives the cost, and what is always included in the price.
Read more →Solutions by business function
If the assessment showed the problem is not specific enough yet, start from the team that has it.
Read more →Scored lower than you hoped?
That is the assessment working. Thirty minutes, no charge, and we will tell you which single thing to fix first, including when the answer is that you do not need us at all.