Charge dispute started by an AI agent: identity verified, dispute posted to the bank's systems
Live in sandbox
- Stage
- Front door and handoff
- Moves
- write back
- Customers
- None named yet
Controls enforced
- Identity verification, evaluation before each step, human review on escalation
Evidence produced
- Decision Trace and Audit Trace
Claims
Six evaluator agents assess every invocation (completion risk, compliance evidence, policy interpretation, data trust, audit readiness, exception patterns), and one deterministic orchestrator decides: execute, execute with safeguards, escalate to a human, or block.
Every run produces a Decision Trace (why each AI action was allowed) and every workflow execution an Audit Trace (the chronological governance record).
Any AI agent, including the customer's own, can start a Callvu process over MCP, and the same evaluation and enforcement apply. Sensitive data never goes into an AI chat; must-not-skip steps stay on a process-controlled screen.