I get asked a lot of questions about Copilot. Through my YouTube channel with 120,000 subscribers and my work as a consultant and advisor to large organisations, I hear from people at every stage of figuring out AI: curious beginners, experienced users, champions, leaders and teams trying to make Copilot work at scale. That gives me a fairly unusual window into what people ask when they're learning in public, what they ask when they're responsible for everyone else, and what they ask when the demo is over and they just want to know why their Copilot can't do the thing they saw five minutes ago. It turns out those questions tell us a lot more about AI at work than I expected, so I started collecting them.

Once you hear enough of these questions, something interesting happens. A thousand different questions start to sound suspiciously like the same handful of questions being asked in hundreds of different ways. People want to know what the AI knows, what it can access, which version they have, which tool they're supposed to use, what it will remember, who else can see their information, whether they can trust the answer, and what any of this is going to cost. They ask product questions, but underneath them are much bigger questions about what it actually means to work with AI.

In this session, I'll share what I've learned from having a front-row seat to so many people figuring this out in real time. We'll unpack what someone might actually mean when they say "Copilot isn't any good," why choosing the right AI is becoming more difficult than writing the prompt, and what happens when the technology is changing faster than people can learn to use it.

This won't be a tour of every Copilot feature. For one thing, we'd need a much longer session. For another, there's a reasonable chance three of them would change while I was presenting. Instead, we'll use real questions from real people, including some wonderfully blunt ones, to work out where people are actually getting stuck and the mental models that make all of this easier to navigate.

Along the way, I'll share the misconceptions I keep hearing, the explanations that seem to make things finally click, and a few things all these questions have changed my mind about.

This session is for business users, leaders, champions and anyone who supports people using Copilot.

You'll leave with a much clearer way to think about AI at work, even when the buttons move, the names change and another shiny new capability turns up tomorrow. And if you've ever wondered whether everyone else understands this better than you do, I can save you the suspense: they don't.