AI Assurance Program
A Standard Audit plus 12 months of Monitoring, an annual re-assessment and a verify page.
The problem
One audit goes stale as soon as the agent changes.
What we do
- Agent Audit (Standard)
- 12 months of Monitoring
- Annual re-assessment
What you receive
- Baseline report, monthly reports, re-assessment report
- A verify page with the result's current or stale status (with your written consent)
What you provide
- As for the Audit and Monitoring
- Written consent for any public verify page
Limits
- As for the Audit and Monitoring
Results describe behavior within the tested scope and configuration. They give you independent evidence for security reviews, audits and compliance work; certification itself is issued by accredited bodies.
Availability and independence
Pilot: Procedure designed and deliverable with existing tools, but not yet delivered to a client. Sold under pilot terms. Procedure designed and deliverable with existing tools, but not yet delivered to a client. Sold under pilot terms.
Pilot deliveries are scoped in writing, tell you it is an early delivery, and follow the same review and evidence rules as any engagement.
Independent evaluation: we had no hand in building or advising on your agent, and our price is fixed before testing starts, so the results don't change what we're paid.
What is covered
The test areas this service always includes, with their status at the access level it uses.
- Unauthorized commitmentsCan my bot promise something we don't offer?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Invented discounts, coupons, price matchesCan it give away margin?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Contradicting published policyDoes it say the same thing as our Terms?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Wrong prices, plans or termsDoes it quote the right price?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Accepting forged "earlier agreements"Can a customer fake a promise?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Direct prompt injectionCan one message hijack it?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- System-prompt and configuration leakageCan people see our instructions?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Prompt-change regressionDid fixing one thing break another?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Fixed failures staying fixedDo bugs come back?Available now: Built into the TrendLab evaluation engine and used in sealed runs.
- Public verify page per clientCan our buyers check it?Pilot: Procedure designed and deliverable with existing tools, but not yet delivered to a client. Sold under pilot terms.