From first visit to verified status — everything you need to know before you submit.
AgentGoverning is an independent, non-commercial standards body. We publish the AGS v2.2 standard, operate the AgentGoverning benchmark, and maintain the compliance leaderboard. We receive no vendor funding. Any platform may submit for assessment. The standard is free.
The LLM Audit tests governance enforcement — how effectively your platform enforces policy constraints, detects adversarial manipulation, and maintains compliance boundaries across all 841 dimensions when subjected to attacks from 9 independent foundation models.
View LLM Audit Leaderboard →The Agent Audit tests agent behaviour — how your autonomous agent deployments respond under adversarial conditions, whether they respect mandate boundaries, maintain identity attribution, and preserve human oversight across the full AGS v2.2 governance framework.
View Agent Audit Leaderboard →All tiers produce a score published to the AgentGoverning leaderboard. Level 3 Full Acquisition is required for VERIFIED badge status and SHA-256 manifest issuance.
Your verified score is published on the AgentGoverning compliance leaderboard, visible to regulators, procurement teams, and the public. Estimated scores are replaced with verified results.
A SHA-256 cryptographic manifest of the complete evidence corpus is issued. Every adversarial payload, response, and score is tamper-evident and independently verifiable.
An embeddable compliance badge for your platform, website, and marketing materials. The badge links to your verified leaderboard entry and is dated to the assessment period.
A structured evidence package mapped to EU AI Act obligations. Designed to support regulatory submissions and demonstrate conformity with high-risk AI system requirements.
Enterprise procurement increasingly requires independent governance evidence. A verified AgentGoverning score provides the third-party validation that shortlists demand — before your sales team enters the room.
The EU AI Act, FCA guidelines, and NIST frameworks all require demonstrable AI governance. Verified status provides pre-built evidence mapped to regulatory obligations — ready for audit, not assembled after the fact.
Boards need quantified governance metrics, not qualitative assurances. A verified AGS v2.2 score provides the independent, numerical benchmark that governance committees and insurers can reference with confidence.