AI Act readiness check

8 yes/no questions map your Annex III + Article 4 + Article 50 exposure. Every question answered "yes" triggers a concrete implication + recommendation. Click to see the implications.

Take 5 minutes and write down each answer. If you answer yes to 1-5: you need M1. If you answer yes to 6-7: you need M2. Yes to 8: both M1 and M2. Yes to several — we recommend M1 + M2 in combination.

1. Do you build an AI system that screens, ranks or recommends candidates in a recruitment flow?

   If yes: Annex III point 4(a) — high-risk classification. The obligations under Art. 16 (provider) and Art. 26 (deployer) apply from the Annex III deadline. Under Regulation (EU) 2026/1744 (Digital Omnibus) the Annex III deadline is 2 December 2027. You should have M1 risk classification + gap analysis in place.

   Affects: Module 1

2. Do you use AI for performance evaluation, promotion recommendations or termination recommendations?

   If yes: Annex III category 4(b) — high-risk. The same Article 6 obligations as recruitment AI. The M1 gap analysis is the minimum documentation.

   Affects: Module 1

3. Does your AI allocate tasks, shifts or projects based on employee characteristics or behaviour?

   If yes: Annex III point 4(b) — high-risk. Task allocation based on behaviour or personal traits is covered by 4(b) — Annex III point 4 only has points (a) and (b). The M1 RACI controls memo + risk classification cover the documentation need.

   Affects: Module 1

4. Do you monitor employee performance or behaviour with AI beyond neutral time tracking (sentiment, productivity, attention)?

   If yes: Annex III point 4(b) — high-risk. Monitoring and evaluating performance and behaviour is covered by 4(b) — Annex III point 4 only has points (a) and (b). M1 covers risk classification + a four-stage approval flow for high-risk decisions.

   Affects: Module 1

5. Do you run any AI systems in production at all (including non-high-risk: chatbots, analytics, generative AI for content)?

   If yes: The Article 4 AI literacy obligation applies (in force since 2 February 2025). You need role-based literacy mapping + an attestation flow. M1 delivers the Article 4 register as documentation.

   Affects: Module 1

6. Do you receive AI-related procurement questions from enterprise customers (SIG, the CAIQ AI addendum, custom buyer questionnaires)?

   If yes: The M2 Procurement Evidence Pack is designed for exactly this: ~145 answers per HR-tech AI system, with Word + Excel + Notion templates ready to hand to your customer. Saves 20-40 hours per customer questionnaire compared with DIY.

   Affects: Module 2

7. Do you need Article 50 transparency disclosure in place (chatbot/AI assistant interaction or AI-generated content)?

   If yes: Article 50 has applied since 2 August 2026. M2 includes chatbot disclosure templates (DA + EN), C2PA AI-content marking notation and a deepfake disclosure flow.

   Affects: Module 2

8. Do you use US-hosted LLM providers (OpenAI, Anthropic, Google) in your AI pipeline?

   If yes: Schrems II + CLOUD Act exposure. M1 documents the data flow + sub-processor list. The M2 Procurement Pack includes SCC citations + Anthropic/OpenAI Schrems II statement summaries.

   Affects: Module 1 + Module 2