Agentic AI is opening doors faster than most organisations can keep up with. With a few clicks, anyone can spin up an AI agent that talks to customers, processes data, or interacts with internal systems – and that speed is both the opportunity and the risk.

In this session, we’ll follow two familiar characters you’ve probably met in your own organisation:

Keith – a renegade Copilot Studio builder. He moves fast, breaks things, and spins up agents quickly to solve real problems but at what cost? Guardrails? “Suggestions.” Best practices? “Optional.” If there’s a limit, Keith has already found a way around it.

Emma – the uncompromising force of Copilot governance. Badge, rulebook, and risk register in hand, she’s here to protect the organisation from data leaks, compliance nightmares, and AI agents that have gone completely off the rails.

We’ll use them to explore the real tension many organisations face right now: balancing creativity and innovation with security, oversight, and responsible use of AI agents living in your very own Copilot city.

This session is grounded in real-world experience from working with organisations who are trying to get Copilot Studio implementation right – not by slowing people down, but by putting the right conversations, boundaries, and support structures in place. By the end, you’ll have a clearer picture of how to encourage responsible innovation, keep things on track, and avoid turning your AI rollout into a firefight.