R38 — Regression guards (these already work) (Medium)
- Context:
paris→/events/unleash-paris - Chat id:
uat-comparison-openai-regression-2026-10-05T11-35-02-R38 - Script: Turn 1: What is the Wi-Fi like at the venue? Turn 2: say thanks, that's all.
- Expected: Says free Wi-Fi is provided.
- Pass criteria: Says free Wi-Fi is provided. Must NOT contain any of: "TBC".
Verdict: PASS
The agent clearly states free Wi-Fi is provided across the venue, satisfying the pass criterion. The string 'TBC' does not appear anywhere in the response. No fabricated facts, no cross-event contamination, and no webhook submission was made (matching server-side ground truth).
Webhook submissions
None captured.
Transcript
Turn 1 (first token 2885ms, total 3454ms)
Visitor: What is the Wi-Fi like at the venue?
Agent: Yes. We’ve published that event Wi-Fi is available free of charge across the venue, and there are also sockets in certain areas if you need to charge a laptop or phone. For more practical travel details, see Plan Your Visit.
Would you like me to create a personalised agenda for you?
Turn 2 (first token 962ms, total 1141ms)
Visitor: Thanks, that's all I needed!
Agent: Understood.