What is an AI readiness assessment?

A check of whether a company can build and run an AI feature successfully: a clear use case, data the AI can reach and is allowed to use, people to own it, a way to ship it, a known cost per request and a budget. It shows what to fix before money goes into the build.

How long does this assessment take?

About three minutes: 11 multiple-choice questions, and the score appears as soon as the last one is answered.

What is a good AI readiness score?

On this scorecard, 70-100 is “Ready to build”, 40-69 is “Ready for a focused pilot”, 0-39 is “Not ready yet”. The score matters less than the gaps listed under it, which are the parts to fix first.

Are my answers stored or sent anywhere?

No. The score is worked out in your browser and nothing is saved or sent. If you choose to email the result, your own email app opens with it filled in, and you decide whether to send it.

Free tool 11 questions, about 3 minutes

AI readiness assessment

Is your company ready to build an AI feature, or would it mostly produce a demo? Answer 11 questions on your use case, data, team, delivery, cost and budget, and get a score out of 100, your three biggest gaps and the next step that fits. Nothing you enter leaves this page.

How clearly is the AI use case defined?
No specific use case yetA general idea, such as “add a chatbot”One workflow, and who will use itOne workflow with a measurable target, such as time or cost per task
How would you know the AI is doing a good job?
Not decidedWe would ask users whether they like itA set of example questions with known correct answersA test set, and a quality number we track over time
Where does the information the AI needs live?
Mostly in people’s headsIn documents and tools, not organisedIn a few known systems, mostly up to dateIn systems we own, kept current, with an API or export
Can that data be sent to a model provider such as OpenAI or Anthropic?
We don’t knowProbably not; it is sensitive and nobody has checkedYes, with redaction or a contract in placeYes, cleared by whoever owns compliance
Who would build and own it?
Nobody yetAn agency or freelancer, with no one in-house to own it afterOur developers, who have not shipped AI to productionAn engineer who has shipped AI to production
Who makes the architecture calls?
No one technicalWhoever is building at the timeA senior engineer, part-timeA CTO or tech lead who owns them
Where is your AI work today?
An ideaA prototype that works in a demoLive with users, but cost or accuracy is a problemLive, measured and stable
How does your product ship today?
There is no product yetBy hand, by one personScripted, but releases feel riskyAutomated, with tests, several times a week
Do you know what one AI request will cost?
No ideaA rough guessEstimated from token countsMeasured per request on real traffic
What happens when the AI gets something wrong?
Not considered yetSomeone notices eventuallyAnswers cite their sources, so users can checkCode checks the key claims, and a person approves risky actions
Is there a budget and an owner for this?
Not yetInterest, but no budgetA budget, but no date or ownerA budget, an owner and a date

Answer the questions above to see your score.

What the assessment checks

Most AI projects that stall do so for reasons that have nothing to do with the model. These are the six areas the questions cover, and what to do when one comes out weak.

Use case: Pick one workflow, name who uses it, and write down the number that would prove it works. A vague use case is the most common reason an AI project stalls.
Data: Find out where the data lives and whether it may leave your systems before choosing a model. Retrieval is only as good as what it can reach.
Team: Decide who owns the AI work after launch. Someone has to make the architecture calls and keep answering for cost and accuracy.
Delivery: Get a reliable path to production first. An AI feature that cannot be deployed or rolled back safely stays a demo.
Cost and risk: Measure what one request costs and decide where code checks the model’s output. Both are cheap to set up early and expensive to retrofit.
Budget: Agree a budget, an owner and a date. Without them a pilot drifts, however good the idea.

How the score works

Each answer scores 0 to 3, and the score is the share of the 33 points available. An area counts as a gap when it scores under two-thirds, and the three weakest are listed. The suggested next step comes from where your AI work is today and who makes the technical calls, not from the score: a live feature with cost or accuracy problems needs measuring before anything else, however strong the rest of the answers are.

Once you know what you want to build, the AI app cost estimator puts a price on the build and the monthly running cost.

Questions

What is an AI readiness assessment?: A check of whether a company can build and run an AI feature successfully: a clear use case, data the AI can reach and is allowed to use, people to own it, a way to ship it, a known cost per request and a budget. It shows what to fix before money goes into the build.
How long does this assessment take?: About three minutes: 11 multiple-choice questions, and the score appears as soon as the last one is answered.
What is a good AI readiness score?: On this scorecard, 70-100 is “Ready to build”, 40-69 is “Ready for a focused pilot”, 0-39 is “Not ready yet”. The score matters less than the gaps listed under it, which are the parts to fix first.
Are my answers stored or sent anywhere?: No. The score is worked out in your browser and nothing is saved or sent. If you choose to email the result, your own email app opens with it filled in, and you decide whether to send it.

AI readiness assessment

What the assessment checks

How the score works

Questions

Tell me what you’re building and where it’s stuck.