The EU AI Act Debugger Tool: 39 everyday business decisions

Classified under Article 5, Annex III and Article 6(3) of the EU AI Act. Runs in your browser, no data stored.

Select which decision situations run automated in your organisation, or are planned to be. The evaluation maps each situation to the EU AI Act – Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744: from prohibited practices under Article 5 through the high-risk area of Annex III to processes that fall outside the Regulation when implemented as rule-based workflows. The check runs entirely in your browser – no cookies, no storage, no external requests.

A – Recruitment and candidate selection

Automation in the intake of new employees – from job-ad targeting to rejection.

B – Performance, behaviour and remuneration

Evaluation and monitoring within existing employment relationships – including variable pay.

C – Tasks, shifts and work organisation

Who works when, where, on what – and on which data basis is that decided?

D – Promotion, probation and termination

The most consequential HR decisions – and the most clearly regulated ones.

E – Training, coaching and development

Access to qualification – partly Annex III No. 3, partly outside the Regulation.

F – Communication, administration and routines

Workflow automation and external communication – here the architecture decides, not the topic.

Result of your selection

0Prohibited since 2 February 2025 (Art. 5)
0High-risk from 2 December 2027 (Annex III)
0Borderline – case-by-case review
0Outside the Regulation

The follow-up question: does the AI sit before or after your decision?

The Regulation requires a human overseeing a running AI process – human in the loop. More effective, and correct in substance, is the reverse order: people decide first, the AI reviews afterwards. AI in the loop means the machine analysis meets an already formed human judgement as a counter-check, instead of pre-formulating that judgement. Whoever thinks first and lets the machine cross-examine afterwards gains both: the speed of the system and the judgement no system brings.

Until a moment ago, you did not know the exact number of your processes running in the high-risk area.
The next step is a review of your specific decision architecture: decision consultation within 48 hours, in person in Frankfurt, Schweizer Straße 5, or online in English. Appointments by phone: +49 69 4800 8888 – payment secures the slot.

EU AI Act Executive Supervision – the full programme

Important notice: This self-check serves general orientation only and does not constitute legal advice. The classifications assume that an AI system within the meaning of Article 3(1) of Regulation (EU) 2024/1689 is deployed; purely rule-based software without inference capability regularly falls outside the Regulation. Article 6(3) provides exemptions from high-risk classification for narrow, preparatory or purely procedural tasks – their application must be assessed and documented case by case. For a legally binding assessment of your individual case, consult a lawyer. Legal status: 9 August 2026, Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744. The check runs 100 % in your browser; no input is stored or transmitted.