EU AI Act
The EU AI Act’s high-risk obligations became enforceable on August 2, 2026: human oversight, logging, and AI disclosure. Our architecture maps onto those duties — and “maps onto” is the claim, chosen carefully. What it does not mean is stated plainly at the bottom of this page.
The mapping
The limit
Using Axioma Facere does not make your business compliant with the EU AI Act. Compliance attaches to your use: what you deploy the roster on, in what role, facing whom. No vendor can sell you that conclusion, and we will not pretend to.
What the architecture gives you is the shape the Act keeps asking for — oversight that is enforced, records that are append-only, disclosure that is structural. If your assessment needs those to exist, they exist here, and you can read exactly how on the governance page.
Direct answers
Your ledger, in your panel: every action attributed, every trust change with its written reason, every cent metered. Append-only is the property that matters — the record of a widening cannot be quietly edited after the fact.
Yes. Disclosure is not configurable politeness here; every name states that it is AI in its external communications. If that costs an occasional sale, we have chosen the cost.
No. It maps our architecture onto the Act’s oversight, logging, and disclosure duties so your counsel has something concrete to assess. The assessment is theirs.
The governance page holds the four hard constraints and the trust ladder — the mechanisms this page mapped onto the Act.