AI doubled your email time and killed 9% of your deep focus

Every tool that promises to save you time eventually generates more of the work it was supposed to eliminate. AI is no exception, but the scale of the backfire is staggering.

AI productivity gains are funding more work, not less

ActivTrak's Productivity Lab analyzed 443 million hours of work activity across 1,111 companies and 163,638 employees. When they isolated 10,584 workers and compared behavior 180 days before and after AI adoption, the results contradicted every efficiency narrative Silicon Valley has been selling.

Time spent on email increased 104%. Chat and messaging usage climbed 145%. Business management tools rose 94%. Not a single activity category showed a decrease after AI adoption. The time AI freed up was immediately repurposed into more tasks, more communication, and more context switching.

As ActivTrak's chief customer officer put it: "It's not that AI doesn't create efficiency. It's that the capacity it frees up immediately gets repurposed into doing other work."

Your focused work sessions are shrinking

The same dataset revealed that the average uninterrupted focus session dropped to just 13 minutes and 7 seconds, a 9% decline. Focus efficiency hit a three-year low of 60%. Meanwhile, collaboration time expanded 34% and multitasking grew 12%.

If you have been wondering why your deep work feels more fragmented lately, this is why. AI tools generate outputs that require review, spark new threads of communication, and create coordination overhead that didn't exist before. The result is a workplace where everyone is busier but fewer people are actually thinking deeply about anything. This pattern mirrors what Berkeley's research on AI workload expansion found: AI doesn't replace effort, it redistributes it.

The 4-tool threshold where everything breaks down

Boston Consulting Group surveyed 1,488 full-time U.S. workers and found a precise tipping point. Productivity climbed with one, two, or three AI tools. Add a fourth, and self-reported productivity collapsed. The researchers coined a term for the fallout: "AI brain fry," defined as mental fatigue from excessive use or oversight of AI tools beyond one's cognitive capacity. Among workers experiencing it, decision fatigue rose 33%, major errors jumped 39%, and intent to quit spiked to 34%. The pattern tracks with BCG's research on AI tool overload, which showed the collapse is not gradual; it is sudden.

The workers most at risk are your best performers

Here is the detail that should worry every manager: 14% of AI-using workers are already experiencing brain fry, and the hardest-hit group is not the reluctant adopters. It is the power users, the ones organizations label "critical to retain."

Marketing roles reported the highest incidence at 26%. Workers under high AI oversight (reading, interpreting, and verifying AI-generated output) reported 14% more mental effort and 19% greater information overload than those whose AI tools operated independently. When managers failed to provide structured support, mental fatigue climbed an additional 5%.

The cruel irony: the employees most enthusiastic about AI are the first to burn out from it. And when they do, AI's rework problem compounds the damage, because correcting AI output eats the very hours AI was supposed to free.

What actually protects focus in an AI-saturated workplace

The ActivTrak data contains one hopeful signal. Workers who spent 7 to 10% of their total work hours using AI tools hit a productivity rate of 95%. The problem? Only 3% of employees currently fall in that range, while 57% use AI for less than 1% of their work time. The sweet spot exists, but almost nobody occupies it.

Organizations that prioritized work-life balance saw 28% lower fatigue scores. Workers whose AI tools handled routine tasks autonomously (instead of requiring constant oversight) reported 15% lower burnout. The distinction matters: AI that eliminates drudgery helps; AI that generates supervision work harms.

If your company just rolled out its seventh AI tool this year (the average organization now uses seven, up from two in 2023), consider applying the 90-minute focus rule before adding an eighth. The data says the next tool will not make you faster. It will make your Saturdays start at 7:11 a.m., which is exactly what ActivTrak found is already happening across their dataset.

---

Related Reading:

• Deep Work Is Failing 68% of Knowledge Workers
• 56% of CEOs got nothing from AI: the $600B Solow Paradox returns