Managing the lifecycle of AI agents

How to handle the AI kill switch without breaking critical workflows

HotTopics Food for Thought In partnership with 

Shutting down an AI agent mid-task is risky business. Pull the plug too quickly, and you risk corrupted workflows, broken downstream dependencies, and real harm to your production environment. But leaving a compromised agent running while teams debate what to do creates its own serious exposure.

The question isn't whether you should intervene—it’s how to do it safely.

At this HotTopics Food for Thought breakfast session, we're bringing together senior technology and security leaders to discuss what a practical agent lifecycle management strategy looks like in the real world. We’ll cover everything from provisioning and monitoring through to suspension, termination, and post-incident review.  

Once the discussion wraps up, we’ll move into an interactive, hands-on cooking session before sharing our creations over a relaxed standing reception.

Event details

📍 Location: Central London

🗓 Date: 12th November 2026

🕛 Time: 9.00 AM - 1.00 PM GMT

• 9:00 AM – Arrivals
• 9.15 AM - 10.45 AM – Breakfast
• 11.00 AM - 12.00 PM – Cooking academy

👥 The Hosts: Doug Drinkwater, Editorial and Strategy Director, HotTopics

Apply to join