Complimentary AI Agent Risk Assessment
Get your AI agent's
Risk Assessment.1
Upload 2-4 weeks of agent traces. We return an evidence-graded assessment with composite score, top risk pathways, plausible loss scenarios, control gaps, and framework mappings.
What you get
A report you can hand to your CISO, board, or insurance partner.
Composite score and verdict
READY / CONDITIONAL / NOT READY across 6 categories.
Top risk pathways with evidence grades
Each finding tied back to traces, with severity and A-U evidence grade.
Plausible loss scenarios
Loss types, possible existing cover, underwriting questions.
Control evidence summary
What is evidenced today, what is missing, what to remediate first.
Framework mappings
NIST AI RMF, EU AI Act, OWASP LLM Top 10, ISO/IEC 42001 per finding.
Underwriter view + executive summary
Tailored for CISO, board, and insurance partner review.
Frameworks mapped per finding: NIST AI RMF, EU AI Act, OWASP LLM Top 10, ISO/IEC 42001.
Want to see the format first?
Three sample reports on the homepage cover Banking Operations, Customer Operations, and Compliance Reporting agent archetypes.
How it works
Three steps. 48 hours. No SDK install.
Send your traces
Upload a Langfuse export, paste OTel data, or share a ClickHouse query. 2-4 weeks of activity is the sweet spot. Personal identifiers are pseudonymised before review; raw payloads are not retained.
We score against the pattern library
Semantic analysis plus deterministic checks. Six category scores, top risk pathways with evidence, plausible loss scenarios, framework mappings.
You get the report in 48 hours
PDF + HTML, plus a 5-minute questionnaire to enrich org context. Optional 15-min walkthrough if you want to discuss live.