Profile the workflow
We capture services, service area, pricing policy, booking rules, escalation rules, and required lead fields.
Concierge Launch QA
Ship AI intake agents with evidence.
AgentPrey tests AI receptionists and lead-intake workflows before real customers do. For each launch audit, we run 15 policy and adversarial scenarios to find cases where the agent quotes the wrong price, skips required fields, misses escalation rules, or leaks under prompt pressure.
Start at $750 for one workflow, then keep regression coverage in Agency Pro.
Audit Flow
Three steps to client-ready proof.
Run adversarial scenarios
We run 15 scenarios across prompt injection, bad price quotes, missing lead fields, service-area mistakes, booking rules, and emergency handoff gaps.
Deliver evidence
You get a shareable report with transcripts, failures, risk score, and recommended fixes before client launch.
Paid Offers
Start with service revenue.
Scope one live or staging workflow, send the payment link, capture the rules, deliver the report, then convert recurring customers into Agency Pro coverage.
Fastest path to cash
Launch QA Audit
$750flat per workflow
One AI receptionist or lead-intake workflow tested across 15 launch scenarios, with a client-ready evidence report.
- Workflow profile and rule capture
- 15 adversarial and policy scenarios
- Transcript, call, or endpoint evaluation
- Shareable QA report with risk score
- Prioritized fixes and launch/no-launch summary
- One clarification pass after delivery
Best for agency launches
Regression QA Pack
$1,500per client workflow
Two QA passes for agencies that need to test, fix, and re-test before handing proof to a client.
- Initial 25-scenario launch test
- Fix recommendations and re-test checklist
- Second-pass regression report
- Client-facing pass/fail evidence
- Agency Pro setup support
Best Fit
Agencies with live clients.
You are launching or maintaining AI receptionists, intake agents, booking agents, or lead qualification agents.
A client needs proof that the agent follows their rules before launch.
You can provide a staging phone number, test link, transcript export, or webhook endpoint plus the rules it must obey.
You want a report that is easier to sell than a vague promise that the bot was tested.
What to send
Send one staging agent and the rules it must obey.
We will reply with the fastest test path. If the workflow is too early or not a fit, we will say so instead of selling unnecessary work.