1. Independence

No carrier, regulator, vendor, law firm, consultancy or other third party pays for placement, ranking, framing, prominence or inclusion in any content on this site. That covers articles, briefings, methodology documents, comparisons and every other form of publication.

Editorial and methodology decisions are made by the editorial and technical staff. Any commercial relationship is kept apart from those decisions and has no influence over what is covered, how it is framed or whether a party is included.

Future Proof Intelligence publishes this site and has a direct commercial interest in the methodology published here, because it intends to offer assessments under it. The interest is disclosed openly. It does not alter the substance of the methodology, which rests on Regulation (EU) 2024/1689, ISO/IEC 42001:2023 and the NIST AI Risk Management Framework, and the methodology is published in full at Methodology v2.0 so that anyone can check the basis of any assessment.

2. What we cover

We cover what carries real significance for the development and use of AI agent certification. The main criteria:

  • A statute, regulation or piece of supervisory guidance creates or changes the obligations an AI agent may be assessed against.
  • A standards body, such as ISO, NIST, CEN or ETSI, publishes or updates a framework relevant to the safety, transparency or governance of AI agents.
  • A court decision, enforcement action or supervisory opinion clarifies the legal weight of certification or of documented AI governance.
  • A carrier, underwriter or institutional investor publishes criteria that refer to certification in coverage or due diligence.
  • An operator, provider or public body publishes assessment results, certification outcomes or audit findings relevant to practice in the field.

We do not cover promotional announcements or vendor claims that have not been checked against primary sources.

3. How we cite

We cite primary sources wherever they exist.

EU legislation by official designation and article, for example Regulation (EU) 2024/1689, Article 9(1), with the recital number where a recital is cited.

Standards by official designation, for example ISO/IEC 42001:2023, and the clause relevant to the claim. We do not paraphrase a requirement without giving its clause.

Court decisions by full case name, court and date of decision, with the ECLI identifier or neutral citation where one exists.

Supervisory guidance by issuing authority, official title and date of publication.

Carrier and underwriter criteria by the institution's own published guidelines or product documentation, with the date.

Trade publications and research by publication or organisation, title and date. We do not treat them as verified fact, and look for the primary source behind them.

4. How we correct errors

We correct errors, and we distinguish between two kinds.

A material error would leave a reader holding a false belief about a regulatory requirement, a provision of a standard or the result of an assessment. It is corrected with a dated note on the affected page, and the original text stays visible, struck through, so the change can be seen.

A non material error is corrected in place, with a note of the date.

Every material correction is logged at corrections. Readers who find an error can tell us through the request page.

5. Conflicts of interest

Future Proof Intelligence publishes this site and five related publications: Future Proof Certified, Agent Liability Europe, Agent Liability Global, Agent Insured and Insure Your Agent. It has a commercial interest in a market for AI agent certification, because it intends to offer assessments and may in time run the partner programme described on the partners page.

These interests are disclosed here and on that page. The methodology is published in full and measured against public standards, so any reader can judge its independence. Where a commercial interest bears directly on a coverage decision, the content says so.

Editorial and methodology staff disclose any personal financial interest in an entity they assess or write about, and such interests are reviewed before publication or assessment.

6. Right of reply

Any organisation named in content or in an assessment on this site has a right of reply. Send it through the request page with "Right of reply" and the content or assessment concerned. We acknowledge it within five working days.

Replies that are substantive and relevant are added as updates or as clearly labelled response notes.

7. Commercial activity

This site operates on an editorial basis for third party coverage. No paid placement or commercial partner listing appears on any page.

A partner programme for verified institutions such as carriers, brokers, consultancies and implementation partners may be introduced in 2027. It has not launched and has no terms. If it is introduced, every paid placement will be labelled as commercial, and partnership will not affect assessment outcomes, certification levels or the framing of any analysis.

8. Versions

These standards are version 1.0, dated 24 April 2026. On 1 October 2026 two factual statements were corrected: a reference to published assessment rates, which do not exist, and a reference to indicative partner pricing, which has been withdrawn. They are reviewed every year, and proposed revisions are published for at least 30 days of public comment before adoption.

9. One standard across the family

The same editorial standards apply across the Future Proof publications. Each publishes its own copy: