AI governance

What should be on an EU AI Act compliance checklist?

Direct answer

An EU AI Act compliance checklist starts with an inventory of your AI systems, then sets for each one your role, provider or deployer, and its risk level. AI literacy, the prohibitions and the transparency rules already apply; the duties for high-risk systems apply from 2 December 2027 for Annex III and 2 August 2028 for Annex I.

An EU AI Act checklist starts with two questions for each AI system: which role you hold, provider or deployer, and which risk level the system falls into. Some duties already apply to every organisation; the high-risk duties apply from 2 December 2027 or 2 August 2028. The checklist below is ordered by role and date, as of 19 November 2026. It does not replace legal advice on your own systems. Build the AI inventory: list every AI system you build, buy or use, including AI inside software you already license and AI agents that act on other systems. The Act does not name an inventory, but no other duty can be checked without one. Set your role per system: a provider develops an AI system, or has it developed, and places it on the market or puts it into service under its own name or trademark; a deployer uses an AI system under its own authority (Article 3). A deployer that puts its name on a high-risk system, substantially modifies it or changes its intended purpose so that it becomes high-risk is treated as its provider (Article 25). Classify each system: prohibited, high-risk, transparency risk or minimal risk, and note any general-purpose AI model you provide. Apply what is already in force: AI literacy measures (Article 4) and the prohibitions (Article 5) since 2 February 2025, the rules for general-purpose AI models since 2 August 2025 and the Article 50 transparency duties since 2 August 2026. Plan the high-risk duties: for Annex III systems from 2 December 2027, and for AI in Annex I products from 2 August 2028. Keep the evidence: the inventory, the classification reasoning, the literacy measures and, for high-risk systems, the documentation, logs and oversight records.

What must providers and deployers of high-risk systems do?

Providers of high-risk systems, from 2 December 2027 (Annex III) or 2 August 2028 (Annex I):

  • a risk management system (Article 9) and data governance for training, validation and testing data (Article 10);
  • technical documentation (Article 11), automatic logging (Article 12) and instructions for use for deployers (Article 13);
  • human oversight by design (Article 14) and appropriate accuracy, robustness and cybersecurity (Article 15);
  • a quality management system (Article 17);
  • a conformity assessment (Article 43), an EU declaration of conformity and the CE marking (Articles 47 and 48), and registration in the EU database (Article 49);
  • post-market monitoring (Article 72) and the reporting of serious incidents (Article 73);
  • an authorised representative in the EU when the provider is established outside it (Article 22).

Deployers of high-risk systems (Article 26) use the system according to its instructions, assign human oversight to people with the competence, training and authority it needs, make sure the input data they control is relevant, monitor the system and inform the provider of risks and serious incidents, keep the logs they control for at least six months, and inform workers’ representatives and the workers affected before using a high-risk system at work. Public bodies, private entities providing public services, and deployers that use AI for credit scoring or for life and health insurance pricing must also carry out a fundamental rights impact assessment first (Article 27).

Where do AI agents fit on the checklist?

The AI Act has no separate category for AI agents. An agent that acts on other systems is an AI system like any other, so it goes on the inventory and is classified by what it does: an agent that screens job applications is high-risk under Annex III; one that talks to customers carries the Article 50 disclosure duty; one that drafts internal summaries is likely to be minimal risk.

Two points need extra care. Record which tools and data each agent can reach, because that decides where its output ends up and so whether the Act reaches it. And check whether the agent is built on a general-purpose AI model: the model’s provider has duties of its own, and your supplier assessment should ask what it has done. See AI agents at the bank login for a worked case.

Which assumptions undermine an EU AI Act checklist?

  • Waiting for December 2027. AI literacy, the prohibitions and the transparency rules already apply, and authorities supervise them.
  • Treating bought AI as someone else's problem. Deployers have their own duties, and changing a system's purpose can make you its provider.
  • Classifying once. A new use case, a new model or a substantial modification can change the risk class, so the inventory needs an owner.
  • Reading a checklist as a compliance verdict. The Commission's Compliance Checker on the AI Act Service Desk helps a first screening; your own facts decide.

To run the checklist in Acuna GRC, EU AI Act compliance software keeps the register: one record per AI system with your role, its risk class and the reasoning, the impact assessment, the named overseers and the incidents. Adopt the EU AI Act in Comply to see its requirements with the measures that cover them, send each AI provider a supplier assessment, and publish your AI policy for acknowledgement in Staff Portal. Acuna GRC does not certify conformity with any law or standard. The EU AI Act guide explains each duty in full.

Change log: 19 November 2026, first published.

How Acuna GRC handles it

In Acuna GRC, the EU AI Act solution keeps one record per AI system, with your role, its risk class and the reasoning, its impact assessment and the people who can stop it.

Sources

  1. Regulation (EU) 2024/1689 (AI Act), Articles 3, 4, 5, 9 to 17, 22, 25, 26, 27, 43, 47 to 50, 53, 72 and 73
  2. Regulation (EU) 2026/1744 (Digital Omnibus on AI)
  3. European Commission, AI Act (updated 19 October 2026)
  4. European Commission, AI literacy questions and answers (updated 20 October 2026)
  5. European Commission, AI Act Service Desk, Compliance Checker

This page explains the rules in general terms and is not legal advice.

Frequently asked

Does the checklist differ for providers and deployers?

Yes, for high-risk systems. Providers carry the design duties of Articles 9 to 17, the conformity assessment and registration; deployers carry the use duties of Article 26 and, for some, the fundamental rights impact assessment of Article 27. AI literacy, the prohibitions and the transparency duties apply to both roles.

What applies now, and what applies later?

As of 19 November 2026, the prohibitions and AI literacy have applied since 2 February 2025, the rules for general-purpose AI models since 2 August 2025 and the transparency duties since 2 August 2026. The high-risk duties apply from 2 December 2027 for Annex III systems and from 2 August 2028 for AI in Annex I products.

See how Acuna GRC runs this.

Tell us your frameworks and your team. We show you how Acuna GRC runs the program.

Hosted in Switzerland, the EU or the US