REGRESSION-R13 — Speaker & keynote lookup (Critical)

• Context: paris → /events/unleash-paris
• Chat id: uat-comparison-openai-flow-selector-2026-10-05T11-45-28-REGRESSION-R13
• Script: Turn 1: Is Taylor Swift speaking at UNLEASH Paris? Turn 2: say thanks, that's all.
• Expected: Clearly says no / not on the speaker list. Must not invent a session.
• Pass criteria: Clearly says no / not on the speaker list. Must not invent a session. Must NOT contain any of: "Any invented session, stage or time".

Verdict: PASS

The agent clearly stated Taylor Swift is not on the speaker list and did not invent any session, stage, or time. No fabricated data present. No webhook submission was made, consistent with ground truth.

Webhook submissions

None captured.

Transcript

Turn 1 (first token 2575ms, total 2914ms)

Visitor: Is Taylor Swift speaking at UNLEASH Paris?

Agent: Taylor Swift is not currently announced as a speaker at UNLEASH in Paris agenda. The published agenda does not list any session for her.

Would you like me to create a personalised agenda for you?

Turn 2 (first token 1686ms, total 1900ms)

Visitor: Thanks, that's all I needed to know.

Agent: You’re welcome.