Every AI system, classified
Each system in your estate gets one verdict — prohibited, high-risk, limited-risk or minimal-risk — with the citation attached.
— weeks until 2 August 2026 enforcement.
EU AI Act · EU
Your EU AI Act evidence, ready before 2 August 2026.
Find every AI system you run. Classify it against the Act. Ship the evidence pack your market-surveillance authority asks for. In days, not months.
What you get
Three outcomes a regulator can use.
The evidence pack
One PDF per system: classification rationale, risk register, human-oversight record and the reviewer signatures the Act asks for.
The incident clock
Serious-incident notifications drafted from the system file, with the deadline surfaced before it runs out.
How it works
Connect. Classify. Ship.
1. Connect
Read-only OAuth into your cloud and SaaS. Metadata only — no model weights, no training data.
2. Classify
Every system mapped to Annex III. Two reviewers sign off before a verdict lands.
3. Ship
One-click evidence pack per system. PDF, CSV, signed audit trail.
Who it’s for
Providers and deployers of high-risk AI in the EU.
Banks running credit-decision models. Fintech running KYC and anti-fraud scoring. HR-tech screening CVs. Insurers pricing policies. Any firm whose AI output reaches a person in the EU.
Article-by-article guidance
Articles 9, 14, 27, 73, 99 — in our docs.
The per-Article walkthrough — risk management, human oversight, fundamental rights impact assessment, incident reporting and the penalty bracket — lives on docs.secruna.com. Send that page to your DPO.
Find your gap.
Before 2 August 2026.
A 30-minute call tells you which of your systems trigger which obligation, and what evidence is missing today.