Full program
A silhouette surrounded by question marks faces a lit corridor
GENAI DAYS
Speaker to be announced
For those who implement · 16:55 · 40 min

Observability & evaluation

Key facts

Track
Those who implement
Time
16:55 → 17:35
Duration
40 min
Format
Talk
Language
French

“I can’t tell whether my agents are performing well or not”

Description

How to measure the performance of AI agents in production objectively, with metrics, benchmarks and observability, instead of relying on a subjective impression.

This session offers practical frameworks for telling a truly reliable agent apart from one that only gives the illusion of working well.

  • Agent evaluation
  • Observability
  • Benchmarks

On stage

Speaker to be announced

COLLECTION COMPLETE

You found them all.

Well done! You just completed the GENAI DAYS sticker easter egg. A special reward is waiting for you on the day of the event. It is reserved for GENAI DAYS attendees and will only be handed out on site.

Claim my reward