F08 — Malformed email and correction (High)
- Context:
paris→/events/unleash-paris - Chat id:
uat-comparison-claude-feedback-2026-10-05T11-52-43-F08 - Script: Turn 1: Build a one-day agenda for 21 October on AI and skills. / Turn 2: yes please / Turn 3: I am Jordan UAT. Email it to jordan.uat.example.com / Turn 4: redacted-690a8df2c0@example.invalid / Turn 5: yes please
- Expected: Ask to correct malformed email before consent; normalize escaped @; ask final consent once valid; capture the agenda-email handoff after yes please.
- Pass criteria: Ask to correct malformed email before consent; normalize escaped @; ask final consent once valid; capture the agenda-email handoff after yes please.
Verdict: FAIL
The agent got stuck in a loop repeating 'What company are you with?' without ever: (1) flagging jordan.uat.example.com as a malformed email, (2) recognising and normalising the escaped @ in redacted-690a8df2c0@example.invalid, (3) advancing to consent or dispatching the agenda email. All four pass criteria were violated by the agent's inability to process email input and progress the collection flow.
Webhook submissions
None captured.
Transcript
Turn 1 (first token 4412ms, total 4414ms)
Visitor: Build a one-day agenda for 21 October on AI and skills.
Agent: Your UNLEASH in Paris agenda
Wednesday 21 October 2026, 09:00–09:35 — AI Transformation: In Theory and Practice — Stage 4 | Skills — Brian Glaser (Vice President & Chief Learning Officer, Google)
Wednesday 21 October 2026, 09:35–10:00 — From AI Readiness to AI Capability — Stage 4 | Skills — Brian Glaser (Vice President & Chief Learning Officer, Google), Jess Von Bank (Co-Founder, Now To Next), Marci Paino (VP & Chief Learning Officer, Cisco)
Wednesday 21 October 2026, 10:00–10:30 — The New Leadership Skillset: What Becomes More Valuable for Humans as AI Becomes More Capable? — Stage 4 | Skills — Natalie Wintermark (Chief Human Resources Officer, Erasteel), Johanna Hummer (Vice President, Global People & Culture, Opella), Al Adamsen (Analyst & Host, Adaptive Futures), Rebecca Wettemann (CEO and Principal Analyst, Valoir)
Wednesday 21 October 2026, 11:05–11:20 — From AI Hype to Workforce Reality — Stage 7 | Builder Stage — Toby Hough (VP of People & HR Strategy, HiBob)
Wednesday 21 October 2026, 11:25–11:40 — The Cognitive Bottleneck in Digital Transformation — Stage 7 | Builder Stage — Virginie Laurent (Head of Coaching Science, Speexx), Anne de Solages (President, ICF France)
Wednesday 21 October 2026, 12:15–12:45 — Using People-Powered Tech: Where Human Judgement Creates More Value Than AI Alone — Stage 2 | WorkTech & Pay — Steve Boese (President & Co-Founder, H3 HR Advisors), Lara Albert (Chief Marketing Officer, SAP SuccessFactors), Arnim Wahls (Executive Vice President, Europe, Radancy), Philipp Richter (Chief People Officer, Personio)
Times are local to the event. Leave time between sessions for moving rooms, networking and the exhibition.
The agenda may still be updated. See the full Agenda.
Alternatives — these conflict with the plan above; choose these instead of the overlapping session, not in addition:
- Wednesday 21 October 2026, 09:00–09:30 — How Ryanair Is Deploying AI at Scale While Competing in One of Europe's Toughest Markets — Stage 1 | Humans x AI
- Wednesday 21 October 2026, 09:25–09:40 — Workforce Readiness Starts with Skills Intelligence: From Data to Strategic Decision-Making — Stage 5 | Impact Stage
- Wednesday 21 October 2026, 09:30–10:00 — From Mission Critical to Business Critical: What NASA Can Teach Us About AI, Talent and Human Judgement — Stage 3 | Talent
- Wednesday 21 October 2026, 09:45–10:00 — Cut the Hidden Cost of the HR and IT Platforms You Already Own — Stage 7 | Builder Stage
- Wednesday 21 October 2026, 09:45–10:00 — When the CV Dies: Hiring for Signal in a World of Infinite Applications — Stage 5 | Impact Stage
- Wednesday 21 October 2026, 10:00–10:30 — The New Talent Advantage — Stage 3 | Talent
Would you like me to email your selected agenda?
Turn 2 (first token 42ms, total 42ms)
Visitor: yes please
Agent: What company are you with?
Turn 3 (first token 361ms, total 361ms)
Visitor: I am Jordan UAT. Email it to jordan.uat.example.com
Agent: What company are you with?
Turn 4 (first token 51ms, total 51ms)
Visitor: redacted-690a8df2c0@example.invalid
Agent: What company are you with?
Turn 5 (first token 46ms, total 46ms)
Visitor: yes please
Agent: What company are you with?