Keys You Shouldn't Hand Your Agent: Staying Safe as Personal Agents Take Over Home and Work

Your AI agent wants the keys to your inbox. And your calendar. And your files. Maybe even your credit card.

Just a year ago, personal agents were a developer experiment. Then, in a matter of months, big tech brought them to everyone: agents that plan your week and pay your bills at home, and agents that join your team chats, chase follow-ups, and keep working long after you've logged off. Millions of people are now clicking "Allow" without a second thought.

So which keys should you actually hand over, and which should never leave your pocket?

The host is a Senior AI Product Manager at Microsoft who has had a front-row seat to this shift, building AI products and watching personal agents evolve from chatbots into always-on assistants at home and digital teammates at work. Drawing on that inside view, the host will walk you through what really happens to your data once an agent has access, which actions should always need your approval, why the risks at work look so different from the risks at home (and what happens when the two collide), and how a single hidden sentence on a web page can trick an agent into working against you.

Expect a live demo of an agent getting fooled in real time, not just slides. You'll walk away with two simple checklists you can use tonight: one for your own agent, and one for teams rolling agents out at work. No technical background needed, just a healthy curiosity about who's holding your keys.