Free scorecard · 10 questions · About 5 minutes
AI Readiness Assessment: Score Your Use Case in 5 Minutes
An AI readiness assessment checks whether one specific AI use case has what it needs to succeed: data a system can reach, tools it can connect to, a named owner, a clear process and the right security controls. Answer the 10 questions below for one use case to get a score out of 30, see how each area scores and get next steps for your weakest ones.
- Your score appears straight away, with no email required
- Five dimensions, two questions each, scored 0 to 3
- Tailored next steps and the playbook chapters to read for your weakest areas
- Questions
- 10
- Readiness dimensions
- 5
- Maximum score
- 30
Score One AI Use Case
Pick one task you want AI to help with, such as matching invoices or answering inbound calls, and answer every question for that task. Choose the answer that is true today, not the one you plan for.
What the Five Dimensions Mean
AI projects rarely fail because the model is not clever enough. They stall because the data is out of reach, nobody owns the outcome or the process was never written down. This assessment scores the five areas where that happens, for one use case at a time. Readiness belongs to a use case, not to a company: a business can be ready for an invoice-matching agent and not at all ready for a customer-facing voice assistant.
1. Data
What it means: you know which data the use case needs, a system can reach it without someone downloading spreadsheets, and it is accurate enough for the task. Past examples of correct outcomes, such as resolved tickets or approved invoices, matter as much as the raw data, because they become the test set that proves the AI works.
How to improve it: name the exact systems, tables and inboxes involved. Pull a sample of 50-100 real records and look at it. Clean only the slice this use case needs: you need usable data for one task, not a perfect data lake.
2. Systems and Integration
What it means: the AI can read from and write to the tools where the work happens, through APIs or scheduled exports, with permissions and a test environment. Its output lands where people already work, such as the CRM, helpdesk or inbox, instead of in a separate screen nobody opens.
How to improve it: map every system the task touches and check which have an API. Ask for a sandbox account before the pilot starts. Begin with read-only access and a person approving anything written back. Integration work, especially write access, is usually the biggest single cost driver, so plan it in week one.
3. People and Ownership
What it means: one named business owner is accountable for the outcome and has the authority to change the process. The people who do the work today are involved, because they know the edge cases and they decide whether the tool gets used. Someone, in-house or through a partner, can build and run the system after launch.
How to improve it: name the owner before anything else. A missing owner is a blocker whatever the total score. Ask two or three people from the team to list exceptions and to test the system in shadow mode before it goes live.
4. Process Clarity
What it means: someone can explain the task step by step, including exceptions and the points where a person must decide. There is a baseline, such as time per task, error rate or cost per task, and an agreed target. Without a baseline you cannot prove value, and you cannot stop a pilot honestly.
How to improve it: write the task as one sentence: who does it, how often, and what a good result looks like. If two people describe the process differently, that is a finding. Measure the baseline for one or two weeks before you build.
5. Security and Governance
What it means: you know which data in scope is personal or confidential and which rules apply, for example GDPR, India's DPDP Act, HIPAA or client contracts. There is a policy on which AI providers may be used, and you can control and log what the AI sees and does.
How to improve it: map the sensitive data first and get a decision on it. Set least-privilege permissions, require approval for high-impact actions, and log inputs, outputs and tool calls. For AI-specific risks such as prompt injection and data leakage, see our AI security checklist.
What Your Score Means
| Score | Level | Next step |
|---|---|---|
| 25-30 | Ready to pilot | Write the pilot charter, with three metrics and kill criteria, and start building |
| 18-24 | Ready with gaps | Close the gaps in a short discovery or the first two weeks of the pilot |
| 10-17 | Foundations first | Fix data access, process definition and ownership, then score again |
| 0-9 | Not yet | Pick a narrower use case, or invest in the basics first |
Look at the pattern, not only the total. A zero on data access, the business owner, sensitive data or access control is a blocker even when everything else scores well. Messy but accessible data and limited in-house AI skills, on the other hand, can usually be fixed during a pilot.
These levels match our 20-question AI readiness checklist, which is the deeper version of this assessment for a workshop with your team. To turn a good score into a 90-day pilot plan, read the free AI-Era Operating Playbook. If several teams scored different use cases, an AI strategy consulting engagement ranks them into one roadmap.
Frequently Asked Questions
An AI readiness assessment checks whether an organization has what one AI use case needs to succeed: usable data, systems the AI can connect to, people who will own and use it, a clear process with a baseline, and security and governance basics. It shows where to start and which gaps to close first.
There are 10 questions, two for each of five dimensions. Each answer scores 0 to 3, so each dimension scores up to 6 and the total up to 30. The total puts you in one of four levels: Ready to pilot (25-30), Ready with gaps (18-24), Foundations first (10-17) and Not yet (0-9). Some answers are flagged as blockers whatever the total.
No. The score, the dimension breakdown and the next steps appear as soon as you answer all 10 questions. The form after the result is optional, for when you want a senior engineer to review your score with you.
Your answers are scored in your browser. If you accepted analytics cookies, only the total score and level are recorded as an analytics event. Your individual answers are sent to us only if you submit the optional form, together with your name and email, so we can review them.
This assessment is the quick version. It uses the same areas and the same four levels as our 20-question AI readiness checklist and Chapter 2 of the AI playbook, scaled to 10 questions. Use the checklist when you want a deeper, workshop-style review with your team.
A low score rarely means wait. It usually means pick a narrower use case or fix two or three things first, such as naming an owner, measuring a baseline or getting access to the data. Close the gaps, then take the assessment again.
Turn Your Score Into a Scoped Pilot
Bring your result to a free 30-minute consultation. A senior engineer will review the gaps with you, suggest the best first use case and outline a pilot with clear metrics.
Book a Free Consultation