Bronze
External Assurance
What can we independently verify from outside?
- — Provider and legal identity
- — Website and external security
- — Privacy and policy analysis
- — AI transparency
- — Automated accessibility checks
AI Trustmark Assurance
Independent assurance for AI products and providers.
Verify the company. Test the operation. Assess the technology.
Evidence-backed AI assurance mapped against relevant controls from leading international AI procurement, security and governance frameworks.
Silver
Operational Assurance
Does the organisation actually operate as claimed?
- — Everything applicable in Bronze
- — Support and recourse testing
- — Policy-to-practice privacy tests
- — Operational evidence
- — Resilience and governance
Gold
Technical Assurance
How is the product itself built, secured and tested?
- — Everything applicable in Silver
- — Source, dependency and secret analysis
- — SBOM, IaC and CI/CD review
- — AI adversarial testing
- — Agent, reliability and technical evals
Assessment depth
Bronze, Silver or Gold
Shows how much evidence was examined and how deeply the product was tested.
Assurance Score
0–100
Scores applicable controls. A high score cannot compensate for a catastrophic hard-gate failure.
Product risk
Low, Elevated or High Impact
Determines applicable controls, evidence strength, testing intensity and certification gates.
Procurement ready
One control catalogue. Multiple framework mappings.
Controls are versioned once and mapped to relevant requirements from the UK Government AI Playbook, PPN 017, NCSC secure AI guidance, NIST AI RMF and Generative AI Profile, applicable EU requirements, US federal acquisition guidance, and government assurance frameworks from Canada, Australia and Singapore.
Reports state how many applicable mapped controls were independently evidenced. They do not claim government approval, NIST certification, blanket legal compliance or EU AI Act certification.