The AI Conference for Experimentation Leaders the humans in the loop

AI can summarise a thousand test results. It can't sit across from someone who already solved your exact problem. Moderators run the small-table conversations that make Circus work, on topics matched to what attendees tell us they're stuck on.

• 8 peers per table, 45 minutes per round, 1 topic each time.
• You facilitate, you don't present. No slides and no pitching.
• Topics come from the attendee survey, so you'll be matched.
• Moderators get a free conference ticket and VIP party invite.

The day before the conference is for building. You'll put AI to work on a real experimentation problem, then check its output the way you'd check any test result. A tool you haven't verified isn't a workflow yet.

• Two hands-on sessions, in the morning and afternoon.
• By practitioners who run these workflows in their own day jobs.
• You build it yourself, on your own data where you can.
• An optional add-on to your conference ticket

AI made building cheap. It didn't make deciding what's worth building any easier. Our speakers run experimentation inside real companies, and they'll show you where the machines help and where a human still has to make the call.

• Talks from in-house practitioners, not sponsors buying stage.
• Real systems and real results, including the parts that broke.
• Talks sit between table sessions. You can pull an ideas apart.
• Lineup drops on a rolling basis.