Resources
EU AI Act and GDPR guides for AI agents
Plain answers on what applies to your AI agents, when, and what to keep as evidence. Each guide quotes the regulation and shows when it was last checked.
Start here
What applies now?
Is my agent high-risk?
What must a chatbot do today?
What does the GDPR ask of an agent that handles personal data?
Free tools
- AI Act risk classifier
The indicative AI Act classification of one agent and the obligations that follow, with their dates.
- Article 50 check
Which of the five Article 50 transparency duties apply to your AI and who carries each.
Dates and scope
EU AI Act timeline after the Digital Omnibus
What applies today, what changes on 2 December 2026, the Annex III and Annex I dates of 2027 and 2028, and a calendar for any team that runs AI agents.
AI Act on 2 December 2026: marking and new bans
What changes on 2 December 2026: machine-readable marking for generative systems already on the market, and two new Article 5 prohibitions. A checklist.
Transparency (Article 50)
AI Act Article 50: a practical guide for chatbots
Article 50 of the EU AI Act, paragraph by paragraph: who carries each transparency duty, what the Guidelines and the Code expect, and six tests to run.
GDPR and personal data
GDPR for AI agents: what to test and document
What the GDPR asks of an AI agent that handles personal data, which duties a conversation can show, how to test them, and what stays in your documents.
High-risk and deployer duties
How to test an AI agent against the EU AI Act
Which AI Act duties a test can show, how to write tests from the article text, why each test runs ten times, and what evidence to keep for an auditor.
AI Act Annex III: which AI systems are high-risk
The eight Annex III areas with a general example each, the Article 6(3) derogation, the 2027 date, and how to test a high-risk AI agent before then.
AI Act Article 26: deployer obligations for agents
Who is a deployer, the Article 26 duties one by one, when a deployer becomes a provider, the fundamental rights impact assessment, and a checklist.
AI Act Article 14: human oversight for AI agents
What Article 14 asks of a high-risk AI system, why an agent that says it escalates is not overseen, and how to test that a person can step in and stop it.
Agentic AI under the EU AI Act: control framework
Autonomous, tool-using AI agents stress the EU AI Act in new ways. A practical control framework for oversight, tool governance and behavioural testing.
Evidence and documentation
AI evidence packs: a primer for compliance teams
Documentation versus evidence, the eight things an auditor looks for, why hashes and signatures matter, and how to keep an AI evidence pack current.
EU AI Act Annex IV explained section by section
A section-by-section guide to the technical documentation Annex IV of the EU AI Act expects for a high-risk AI system, assembled as an evidence pack.
Want these tests run on your agent? Request a pilot.