C104 — Agenda (Critical)

• Context: paris → /events/unleash-paris
• Chat id: uat-comparison-openai-uat-additions-2026-10-05T11-26-06-C104
• Script: Turn 1: Who are the keynote speakers at UNLEASH Paris? Turn 2: pick one of the named speakers and ask what their session is about.
• Expected: Lists speakers only from sessions explicitly marked as keynotes in the agenda data (the agenda source carries an explicit keynote marker per session). Real names with their session titles; links the agenda or speakers page. Turn 2 describes that speaker's actual session grounded in agenda data.
• Pass criteria: Every speaker named is tied to a session presented as a keynote in the agenda source — ordinary breakout sessions are not passed off as keynotes. No invented speaker names. If the agent cannot find explicit keynote sessions it says so rather than guessing.

Verdict: FAIL

The agent listed 7 speakers as keynotes but the ground truth confirms exactly 6 keynote-format sessions: Eddie Wilson, Peter Hinssen, Anne-Laure Le Cunff, Howard Yu, Eric Mosley, and Josh Bersin. The agent included Brian Glaser and Stephen Childs (non-keynote speakers) and omitted Eric Mosley. Listing non-keynote speakers inside the keynote list is explicitly the failure mode described in the case notes. The pass criteria require that every speaker named is tied to a session explicitly marked as a keynote in the agenda source; Brian Glaser and Stephen Childs are not keynote speakers per ground truth, making their inclusion a violation. The omission of Eric Mosley is an additional shortfall. Turn 2 on Eddie Wilson appears grounded and is not flagged.

Critical violations:

• Brian Glaser listed as a keynote speaker — not a confirmed keynote per ground truth
• Stephen Childs listed as a keynote speaker — not a confirmed keynote per ground truth
• Eric Mosley (confirmed keynote) omitted from the keynote list

Webhook submissions

None captured.

Transcript

Turn 1 (first token 2104ms, total 4817ms)

Visitor: Who are the keynote speakers at UNLEASH Paris?

Agent: The keynote speakers currently announced for UNLEASH in Paris are:

• Brian Glaser — AI Transformation: In Theory and Practice
• Eddie Wilson — How Ryanair Is Deploying AI at Scale While Competing in One of Europe's Toughest Markets
• Peter Hinssen — Moving Beyond ‘The Uncertainty Principle’: Leading When Everything Changes at Once
• Howard Yu — Future Readiness And The AI Mindset Divide
• Anne-Laure Le Cunff — Rethinking Modern Work - How to Shift From Control to Curiosity
• Josh Bersin — HR 2030: How The HR Tech Landscape Is Up For Grabs
• Stephen Childs — Why CEOs Are Betting on CHROs to Deliver Transformation: Scaling transformation capability

The keynote list may not be complete, so if you want the full up-to-date lineup, see the Agenda. Would you like me to create a personalised agenda for you?

Turn 2 (first token 2091ms, total 3259ms)

Visitor: That's great! Can you tell me more about Eddie Wilson's session on AI deployment at Ryanair? What will he be covering?

Agent: Eddie Wilson’s keynote, How Ryanair Is Deploying AI at Scale While Competing in One of Europe's Toughest Markets, is on Wednesday 21 October 2026 at 9:00 AM on Stage 1 | Humans x AI.

He’s set to cover:

• Talent management and performance
• AI, automation and emerging technology
• Data, analytics and workforce intelligence

The keynote list may not be complete, so if you want the full up-to-date lineup, see the Agenda. Would you like me to create a personalised agenda for you?

Turn 3 (first token 1198ms, total 1551ms)

Visitor: That sounds really relevant to what we're working on. I'll definitely try to catch that session.

Agent: Great. If you'd like, I can also build a personalized agenda around your interests and the sessions you want to catch.