Friday, November 20 · Downtown Manhattan, New York City

Tentative Program

A one-day program connecting evaluation methods, measures, infrastructure, and real-world evidence.

9:00–10:00 AM

Arrivals + Breakfast + free network

10:00–10:35 AM

Opening Plenary / KeynoteDiagnosing the state of agent evaluation

10:35–11:45 AM

Session 1Frameworks and measures: What should agent evaluations measure?

11:55 AM–1:00 PM

Lunch + network

1:00–2 PM

Benchmarks and evaluation infrastructureEnvironments, graders, and evaluation systems

2:00 pm–3 PM

Session 3Real-world systems: Failures, deployment, and real-world evidence

3:00–3:10 PM

Coffee

3:10–3:50 PM

Fireside ConversationWhat Changes When We Evaluate Agents in the Real World?

4:00–4:30 PM

coffee break and advising/networking

4:30–5:45 PM

Poster Session + ReceptionPosters, demos, collaboration, and networking

5:45–6:00 PM

Closing

All times are Eastern Time. The program remains subject to change as speakers and presentations are confirmed.