Here is the finding from data published this week: rolling out AI did not let teams work shorter days. It did the same thing every tool before it did. The number of things people juggle went up, and focus time went down. That makes protecting focus time one of the quietly important jobs a leader has right now, because the default direction of AI is a denser, more fragmented day.

According to the ActivTrak data, focus efficiency fell to a three-year low of 60%, the typical uninterrupted work session shrank to just 13 minutes, and weekend work climbed more than 40%. Thirteen minutes.

This is not an article about AI. AI is a tool, and any tool used carelessly gives a bad result. Try to unscrew something with a pair of scissors, held by the blade, and see how it goes. This is an article about leadership and company culture. If you lead a team, the question is simple: do you value your team's focus time, or not?

Why "faster" quietly becomes "more"

When a task that took an hour now takes five minutes, the time you saved does not stay saved. It gets absorbed. You take on the next thing. You add work you could not have justified before. You raise the bar on what fits in a day. AI lowers the cost of producing work, and when the cost of anything drops, people consume more of it. Output goes up, pace goes up, and the number of things you are juggling goes up right along with them.

That is the mechanism behind the ActivTrak numbers. Collaboration rose 34%. Multitasking rose 12%. People are not busier by accident. They are doing a larger variety of things, faster, which means more switching between them.

The hidden tax: what breaks every time focus does

The reason a fragmented day is so expensive is not the interruption itself. It is the recovery. A landmark University of California, Irvine study by Gloria Mark found it takes an average of 23 minutes and 15 seconds to fully refocus after an interruption. And it is not just lost time, it is degraded quality: a piece of your attention stays stuck on the previous task even after you switch, an effect researchers call attention residue.

Now layer that onto how the day already runs. Microsoft's Work Trend Index found the average knowledge worker is interrupted every two minutes by a meeting, message, or notification, roughly 275 times a day. Forty percent never get even 30 uninterrupted minutes in a workday. If it takes 23 minutes to refocus and you are interrupted every few minutes, deep work never actually happens.

We've watched this movie twice: Slack and Teams

If this pattern feels familiar, it should. Slack and Microsoft Teams were sold as tools to streamline work and cut email. Used well, they do. But most organizations never decided how to use them. They just switched them on, and for a lot of teams the result is an always-on stream of pings that shreds focus worse than the email it replaced. The average worker now gets 153 Teams messages a weekday. The tool is not the problem. The absence of any shared agreement on how to use it is.

AI is the next instance of this exact pattern, only with more horsepower. A powerful tool amplifies whatever culture you already have. Deploy it into a chaotic, interrupt-driven culture and you get faster chaos. Deploy it into a culture that protects focus and you get more done with focus intact.

Density is a design choice, not a destiny

A tool's value comes from how it is used, not from what it can do. Hand your team powerful AI with no guidance and the natural gravity is toward "more": more output, more speed, more things in flight, more interruptions to keep up with it all. That is not a hypothetical, it is what the ActivTrak data shows happening right now across hundreds of organizations. Left to default, AI makes the day denser. But if you explicitly decide that some of the time AI frees up gets reinvested in focus rather than volume, you get a different company.

What protecting focus actually looks like

This is not about banning tools or slowing down. It is about designing how your team works so the speed AI provides is not spent entirely on fragmentation. A few concrete moves:

None of these are new ideas. They are good leadership decisions that build a strong team and support the long-term health of the organization. It is the same discipline that separated the companies that thrived on team chat from the ones that drowned in it. The tools will always change. The principles do not.

AI genuinely is faster. Whether "faster" turns into a calmer, sharper team or a busier, more scattered one is not up to the technology. Design the way of working first, then add the tool. It is up to you.