On August 18, 2026, Lexic.AI signed Pillar II of the European Commission's AI Pact and was confirmed as a registered signatory. It's a small announcement with an unusual detail attached: we're a company that independently audits other companies' AI agents, and we just made a public, voluntary commitment to be governed by the same standard we apply to our clients.
What the AI Pact actually is
The AI Pact is a voluntary framework run by the European Commission's AI Office, launched ahead of the EU AI Act's phased entry into force. It has two pillars: Pillar I gathers organizations to share implementation practices; Pillar II asks signatories to commit publicly, in writing, to specific governance actions ahead of their legal deadline. It is not a certification, and the Commission says so explicitly — signing the Pact is a disclosed intention, not a stamp of approval on a product.
What Lexic.AI committed to
Our signed commitment covers action across the lifecycle of an AI system, not just at launch:
- Develop: identify risks across the system's lifecycle, apply documented data-quality policies for training, validation and testing data, keep logs traceable, tell deployers how the system is meant to be used and where its limits are, keep a human in the loop, and design the system to disclose it is AI to the people who interact with it.
- Deploy: map how the system's risks touch people's fundamental rights, and make sure anyone interacting with it is told they're talking to AI.
That last point isn't new to us — it's Article 50 of the EU AI Act, in force since August 2, 2026, and something Lexic Compass has been testing for in client audits since before that date. Signing the Pact means we hold ourselves to the same disclosure standard we score other companies against.
Why this matters more than it looks like it should
The Digital Omnibus, adopted this summer, pushed the EU AI Act's high-risk obligations (Annex III) out to December 2027. A lot of the market read that as "we have time." What actually happened is narrower: transparency obligations under Article 50 never moved, and the buyers who matter most — banks, insurers, and other regulated enterprises — aren't waiting for December 2027 to start asking AI vendors what they've actually committed to, on the record, today.
An independent, public, dated commitment is a different kind of evidence than a claim on a slide. It's checkable. Anyone can look up the Commission's list of Pillar II signatories and find us on it.
What this isn't
We want to be precise about this, because vague claims are exactly what an AI audit company shouldn't make. Signing the AI Pact is not a certification, an endorsement, or a legal safe harbor. Lexic is not a notified body under the EU AI Act, and neither the AI Pact nor Lexic Compass issues certifications — Compass closes an audit with an independent verdict (Cleared, Cleared with Conditions, Not Cleared), never a certificate. The Commission's own disclaimer, which we carry on our site, says it plainly: this "does not constitute an endorsement or certification of Lexic.AI's products by the European Commission."
Frequently asked questions
What is the European Commission's AI Pact?
A voluntary framework from the EU AI Office that invites organizations to commit publicly to AI governance practices ahead of their legal deadlines under the EU AI Act. Pillar I is about sharing practices; Pillar II is about signing a specific, written commitment.
Is signing the AI Pact the same as being certified?
No. The Commission is explicit that Pillar II signature is not a certification or an endorsement of any signatory's products. It is a disclosed, voluntary intention to act — verifiable, but not a compliance guarantee.
Does this replace Lexic.AI's obligations under the EU AI Act?
No. It sits alongside them. Lexic.AI still has to meet its own obligations as an AI provider (including Article 50 disclosure in our own products), and Lexic Compass still audits other companies' agents against the same regulatory framework, independently of any Pact commitment.
Why would an AI audit company sign this instead of just auditing others?
Because the alternative — asking clients to prove governance we haven't disclosed ourselves — doesn't hold up. An audit is only as credible as the auditor's own transparency.
Where can I verify this?
The European Commission publishes its list of AI Pact signatories at https://digital-strategy.ec.europa.eu/en/policies/ai-pact. Lexic.AI's commitment is also disclosed in the footer of every page on lexic.ai.
If you want to know where your own AI agents stand against the same standard, book a demo with your own data.