Is this checklist the same as being compliant?

No. It is a control inventory that shows which obligations apply and whether you can evidence them. Conformity is the deployer's responsibility across the full system and requires more than a completed checklist.

What is the most-missed item on the list?

A named human-oversight owner (Art. 14). Teams ship the model but never assign the person who monitors and can interrupt it, so the control is unowned in practice.

When should I re-run the checklist?

After any material change — a new model, a new training-data source, a new use case, or a new target market. Material changes can move the risk tier and reopen gaps you thought were closed.

Does the checklist cover the 2 Dec 2027 deadline?

The checklist tracks which obligations apply; the current effective date for many high-risk duties is 2 December 2027 under the Digital Omnibus, but verify the live timeline with counsel before relying on it.

Is ClauseGuard a legal opinion?

No. ClauseGuard returns a risk-tier classification and a gap report. It is decision-support, not a legal opinion, a certification, or a guarantee of conformity.

EU AI Act readiness checklist

---

Before you launch

Readiness is not a single switch. Before you put an AI feature in front of EU users, run a control inventory that tells you which obligations apply and which you can already evidence. This checklist is deliberately framed as "what to check," not "you pass." If any item is unowned, that is your remediation list.

The checklist

Classify the risk tier — Is the use listed in Annex III (recruitment, credit scoring, safety, biometrics, and more)? If yes, high-risk duties (Art. 9–17) attach. If not, check prohibited practices (Art. 5) and transparency duties (Art. 50).
Document the gaps to obligations — For each applicable article (Art. 9 risk management, 10 data governance, 11 technical documentation, 12 logging, 13 transparency, 14 oversight, 15 robustness), note whether you can evidence the control today.
Assign a human-oversight owner (Art. 14) — Name the natural person who understands, monitors, and can interrupt the system. "The team" is not an owner.
Log the assessment — Keep the classification, the gap list, and the evidence artifacts together so they can be produced for a conformity assessment.
Re-run after any material change — A new model, new data source, or new use case can change the tier and reopen closed gaps.

How to use the output

Each unchecked item is a remediation task with an owner and a due date. The finished checklist is an artifact for your conformity file — proof that you looked, classified, and assigned, even where controls are still in progress. Teams that score the system first (e.g., with ClauseGuard) get the gap list generated, then attach owners in this checklist.

What the checklist is not

It is a control inventory, not a pass/fail certification. Ticking every box does not make a regulator declare you compliant; leaving some open does not make you "non-compliant" by itself. Conformity depends on the full deployed system and is the deploying organization's responsibility. Use the checklist to drive evidence, not to claim a status you cannot support.

Authoritative references

EU AI Act (Reg. 2024/1689), Art. 5 / 9 / 10 / 11 / 12 / 13 / 14 / 15 / 50: https://eur-lex.europa.eu/eli/reg/2024/1689/oj
AI Act explorer: https://artificialintelligenceact.eu/
European Commission AI policy: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai