EU AI Act Compliance Inspector

EU AI Act Compliance Inspector

When to Use / Best For

You need to know if a process is legal under the EU AI Act: the skill maps any department, workflow, or tool to the Act's four risk tiers (Prohibited, High, Limited, Minimal) and tells you the legal status of each.
You suspect there's "shadow AI" in your stack: use it when you rely on everyday tools (ATS résumé screening, CRM lead scoring, HR performance monitoring, credit scoring, chatbots) and aren't sure which ones quietly count as regulated AI.
You're racing the 2 August 2026 deadline: the skill flags exactly which obligations apply, when they kick in, and what to fix first, so you act before enforcement, not after.
You want a fast first-pass audit before involving counsel: generate a structured report in minutes instead of waiting on a full consulting engagement, then take it to your lawyer or DPO.
You need a ready-to-share compliance report: the result is a finished, structured audit document you can review, circulate internally, and act on.

How It Works
Step 1 — Tell it what to audit: start with a department, a workflow, or a list of tools. If you only know the department ("audit HR"), that's enough to begin.
Step 2 — Add context if you have it: mention what each system decides (especially decisions about people), what personal data it touches, and whether you build the system (provider) or use a vendor's (deployer). The more you give, the more definitive the verdict.
Step 3 — Get guided when you're unsure: if the input is thin, the skill runs a short discovery round, showing you where AI typically hides in your departments and a quick checklist of what to provide, instead of guessing silently.
Step 4 — Receive the audit report: you get a structured report with a risk-classification matrix, a per-system legal impact analysis, a mandatory compliance shielding protocol (human oversight, audit logging, transparency steps), and a final auditor's verdict with the binding penalty tier.
Example
Input
A company audits its HR department and says:

Audit our HR — we use Workday Recruiting to screen and rank candidates and an Intercom chatbot for applicants.

Output
The skill returns a structured audit report with:

A risk-classification matrix tagging Workday as High-risk recruitment and the chatbot as Limited-risk
The exact legal basis for each (Annex III §4(a); Art. 50 transparency)
The bias and GDPR exposure for each system
A shielding protocol: human-in-the-loop sign-off before any rejection, audit logging, AI-disclosure notice
A final verdict with the applicable fine tier under Art. 99

The exact findings depend on your tools and how they're used, but the purpose is clear: know your exposure and what to fix before regulators or fines arrive.

Why This Skill / What Makes It Different

Grounded in the actual law, not invented: every article, Annex, and penalty figure is anchored to the real text of Regulation (EU) 2024/1689. No fabricated citations, no made-up fines.
Classifies by use, not brand name: the same tool can be Minimal in one workflow and High-risk in another. The skill judges the intended purpose, the way the regulation actually does, instead of guessing from the product name.
Detects hidden AI you didn't flag: it actively infers the algorithmic features embedded in common business software, surfacing regulated AI you may not have known you were running.
Honest about certainty: when you give little, it issues a clearly labeled PROVISIONAL report and tells you exactly what to add for a DEFINITIVE one, instead of faking confidence.
Ends with action, not just a verdict: every report includes a concrete shielding protocol (human oversight breakpoints, immutable logging, transparency measures) to close the gaps before the deadline.

Who It's For

Built for compliance officers, Data Protection Officers (DPOs), founders, legal and GRC teams, consultants, and operations leaders who need to understand and reduce their EU AI Act exposure across HR, finance, customer service, security, and beyond.

Best Use Cases
Use this skill when you want to:

Classify a tool or workflow under the EU AI Act's four risk tiers
Detect shadow AI hidden inside commercial software
Check whether a process is Prohibited or High-risk before launching it
Prepare a compliance pre-audit before involving lawyers
Map your exposure ahead of the 2 August 2026 deadline
Get a clear, structured report with a concrete fix-it protocol

This skill is an automated compliance-screening aid based on Regulation (EU) 2024/1689. It does not constitute legal advice and does not replace review by qualified counsel or a notified body.