An assistant offers to take care of something for you. It can gather information, use tools and take the next step. You may already be working with an AI agent without calling it one.

This session makes that relationship visible for a non-technical audience. A live example lets us follow a task from the original request to the result. Along the way we examine what the agent can see, which instructions guide it, what it can actually do and where it needs a person’s judgement.

A missing attachment, an unclear rule and an action that was never checked lead to different problems. Understanding those differences helps us decide how much trust to give the assistant and what to verify.

Participants leave with a way to recognise when software is acting on their behalf and to ask useful questions about the work they have handed over.

For general audiences curious about AI agents in everyday life and work. A live example makes an otherwise hidden process visible and gives the room concrete decisions to discuss. Participants need no programming background. This is the accessible counterpart to “Congratulations, AI made you the boss”, focused on recognising agency and calibrating trust rather than managing a technical team.