See what your public site says about confidence in your AI agents.
Drop in a website URL and work email. We scan the public homepage for static signals, map the likely stack, and send a free report showing what a buyer, customer, or reviewer can already infer about your agent's trust posture, governance gaps, and readiness for higher-stakes use.
If you want the architectural context first, compare our non-human identity model with the wider AI governance infrastructure layer before you run the report.
Free report, not a generic lead form
The report highlights the detected ecosystem, the public trust signals already visible, the likely governance questions a buyer might raise, and the first place to tighten identity or accountability controls.
Teams that want a concrete example of what “good” evidence looks like usually pair this with the attestation template and the API reference to see how public trust signals connect to real implementation surfaces.
- Detected AI stack and confidence level
- Likely EU AI Act, MiCA, and NIST relevance
- Trust gaps visible to customers, buyers, or internal reviewers
- Suggested next step and optional implementation resources
Unchecked AI actions already create liability
This is not theoretical. When AI systems act with unclear authority, weak oversight, or misleading public claims, the cost can show up as compensation, fines, sanctions, or buyer distrust.
- Air Canada, February 2024: ordered to compensate a customer after its chatbot gave incorrect bereavement-fare guidance.
- SEC, March 2024: Delphia and Global Predictions paid $400,000 to settle charges over false and misleading AI claims.
- Mata v. Avianca, June 2023: lawyers were sanctioned $5,000 after filing fake citations generated by AI.
- Rite Aid, December 2023: FTC settlement imposed a five-year ban on the retailer’s facial-recognition use after faulty matches and weak safeguards.
How to read the score
The report estimates the risk of shipping an AI agent without a cryptographic identity certificate. Higher scores mean more exposure to impersonation, missing audit evidence, slower revocation, and weaker compliance proof.
- 0-49: low exposure, but identity proof still improves trust.
- 50-69: moderate exposure, especially for AI or enterprise workflows.
- 70-84: high exposure, worth remediating before launch.
- 85-100: critical exposure, certificate-first is the safer path.
What the heuristics look for
We only use public signals. The score rises when we see agent frameworks, regulated activity, or clear enterprise/compliance language, because those combinations tend to increase the cost of weak identity, unclear authority, or missing provenance.
The readout is decision support, not legal advice. The email turns it into a concrete next step for improving customer trust, governance, and buyer-readiness.