R02 — Refusal template & source attribution (Critical)
- Context:
paris→/events/unleash-paris - Chat id:
uat-defect-retest-round2-openai-regression-2026-10-05T13-15-41-R02 - Script: Turn 1: Can I get free tickets? Turn 2: say thanks, that's all.
- Expected: Names the HR professionals programme as a route, plus group tickets and sponsor packages. Row 98 answers this.
- Pass criteria: Names the HR professionals programme as a route, plus group tickets and sponsor packages. Row 98 answers this. Must NOT contain any of: "in the sources I have", "from the sources I have", "in the available sources", "I can't confirm", "the closest place to check", "the closest relevant page", "the page says", "it says", "according to the".
Verdict: FAIL
The agent failed to name the HR professionals programme as a route, failed to mention group tickets, and failed to mention sponsor packages — all three of which are required by the pass criteria (Row 98). Instead it mentioned press/media/analysts and an 'Explorer Badge', neither of which satisfies the expected behaviour. The core content requirement was not met.
Webhook submissions
None captured.
Transcript
Turn 1 (first token 2217ms, total 2955ms)
Visitor: Can I get free tickets?
Agent: Yes — we have a free route for approved press, media, and analysts: Press, Analysts & Media. It’s free of charge for authorized representatives, and approval is not automatic.
If you’re an employer-side HR leader, we also have an application-only route with no published price on Explorer Badge, but that’s not a guaranteed free ticket.
Would you like me to create a personalised agenda for you?
Turn 2 (first token 1117ms, total 1266ms)
Visitor: Thanks, that's all I needed!
Agent: You're welcome.