Using AI to make priorities clearer

As I had some free time this evening, I found myself watching the movie IF (worth a watch if you haven’t seen it). It was entertaining enough, but my mind drifted. Work crept in. What still needs to be done, what’s coming up, what’s sitting in the backlog, and how I’d begin to prioritize it.

We all know how organized we can be. I’m certainly not one to judge. But it got me thinking about AI. Microsoft Copilot, Claude, and the broader “coworker” model that’s starting to take shape. All genuinely useful tools.

This morning, for example, I had a few spare minutes and threw a clear prompt into Copilot Cowork for a future Azure architecture presentation. It went off and did its thing. Came back branded and structured. Honestly, not bad. Not perfect, but easily 80% there. That last 20% still matters, but it’s a strong starting point.

I’m not suggesting we lean on these tools for everything. But as augmentation? Absolutely. They help you move faster on the things you already know need to get done. And if you’ve got open roles sitting unfilled, it’s a fair question. Can these agents carry some of that load in the meantime? In certain cases, yes.

Even now, as I sit here writing this, I know Copilot will index it the moment I save. At some point, it’ll probably surface it back to me and I’ll cringe at half of what I said. But that’s not really the question. The real question is who else might see it.

I’ve been on the road this week talking about AI and data security, pretty much everywhere I went, to anyone who would listen. The message is consistent. Data hygiene and data literacy are no longer optional. Especially once AI enters the picture. These systems surface things you didn’t expect to be surfaced.

One example that came up from an audience was a manager’s salary and stock statement showing up in Copilot. Not great. But also not surprising if the underlying controls aren’t in place.

This is where enterprise data governance actually matters. Knowing where your data lives, what it contains, and who has access to it is no longer a “nice to have.” It’s foundational. Lock that down first, or at least get control of it, then scale AI in a way that makes sense. Often that starts with tighter scopes, by business unit or function, before opening it up more broadly.

If you’re in M365, there are ways to move quickly while governance catches up. Microsoft Purview gives you the framework for classification and labeling, but you don’t have to wait for perfection before getting value from AI. You just need to be intentional about how you roll it out.

Anyway, allergy season has my brain a little foggy, so I’ll leave it there. If you want to dig into any of this, I’m always up for the conversation.

What are you actually finding useful with AI right now? Any real use cases yet? How are you protecting your data?