About the event
AI agents take on real work inside businesses but handing over more responsibility also raises some questions: what should an agent be allowed to access? Who is accountable when something is wrong? And how do you know what it has done? For growing businesses, scattered content infrastructure, legacy systems and unclear permissions can quickly become a blocker. Simultaneously, locking everything down can make AI too cumbersome to be useful. This Sifted Talks will look at how companies can find the balance: putting practical guardrails around AI agents without slowing teams down.
Date & Venue
Wed 28 Oct 2026 / 12:00 (GMT), 13:00 (CET)
Virtual (Zoom call)
Sponsored by
We'll be discussing
- What is stopping businesses from putting AI agents into real workflows?
- The right level of access: How can companies give agents enough freedom to be useful without opening the door to unnecessary risk?
- What does good oversight look like when agents are making decisions and taking actions on their own?
- How can businesses build security, compliance and trust into AI adoption without slowing teams down?
Moderators & Speakers
Mick McNeil
Founder & CEO @ Deliverance AI
Eléonore Hébrard
Staff Solutions Engineer @ Box