How I score
Two free self-assessments on this site give you a result. This page shows exactly how that result is calculated, what it means and where its limits are. Every rule below is the rule the tool actually runs.
Four principles
AI Readiness Scorecard
Fifteen questions in five dimensions, three questions each. Every answer has five levels, from ad hoc (1 point) to systematic and automated (5 points).
| Dimension | What it asks |
|---|---|
| Data Infrastructure | How organised and accessible your data is, how good its quality is, how much is digital. |
| Process Maturity | How standardised your key processes are, where the bottleneck sits, how repetitive work is handled. |
| Team & Skills | How familiar the team is with AI, whether technical capability exists, how leadership views AI. |
| Compliance & Validation | How new technology meets regulatory requirements, how ready you are for AI data privacy, whether AI is validated and monitored. |
| Strategic Alignment | Whether an AI strategy exists, whether ROI has been estimated, how AI tools are selected. |
How the score is calculated
- Per dimension: sum of three answers, at most 15 points, shown as a percentage.
- Overall: sum of all fifteen answers, at most 75 points, converted to a score out of 100.
- The lowest possible score is 20, because every answer is worth at least one point.
What the score means
| Score | Profile |
|---|---|
| 20 to 26 | Starting Line |
| 27 to 46 | Foundation Builder |
| 47 to 66 | Growing Momentum |
| 67 to 86 | Acceleration Ready |
| 87 to 100 | AI Frontrunner |
Pattern notes and quick wins
- Balanced profile if your five dimensions lie within 15 points of each other, uneven maturity if they spread 40 points or more.
- Strengths are named from 73 percent in a dimension, a gap at 40 percent or below.
- Quick wins: your two weakest dimensions each get one concrete measure. A third follows from your overall score: one pilot up to 40, a business case template up to 70, an internal champion network above 70.
Limits: the scorecard does not know your company size, budget or product portfolio, and it measures your own view, not an external review. It is built to start the right conversation.
AI Act x MDR/IVDR check
Five questions for medical device and diagnostics manufacturers. They test the route into the high-risk category via Article 6(1) of the AI Act, which covers AI in products under MDR and IVDR.
How the verdict is formed
- The verdict uses three questions only: is it an AI system under Article 3(1), is it a product or safety component under MDR or IVDR, does a notified body take part in the conformity assessment.
- All three yes: very likely high-risk under Article 6(1).
- At least one no: probably not high-risk via Article 6(1).
- Otherwise, with at least one unsure answer: likely in scope, verify.
- Your role (provider or deployer) and your timeline planning do not change the verdict. They decide which hints you get.
Legal basis: Regulation (EU) 2024/1689 as amended by the Digital Omnibus, Regulation (EU) 2026/1744. High-risk obligations for Annex I products, including MDR and IVDR devices, apply from 2 August 2028.
Limits: the check does not cover the second route into high-risk via Article 6(2) and Annex III, for example AI in recruiting, nor transparency duties under Article 50. It does not replace the assessment by your regulatory affairs team.
Changes to this methodology
When the law changes, I adjust questions and texts and record it here. Legal changes themselves are listed in the EU AI Act change log.
| Date | Change |
|---|---|
| 21 Sep 2026 | AI Act check: date for Annex I products corrected to 2 August 2028 after the Digital Omnibus. Methodology published. |
Want to talk through your result?
Bring your score to a free 30-minute call. We look at your weakest dimension and one concrete next step.