Do you know the feeling: reaching the end of a day having accomplished more than ever before, and still feeling exhausted?
That's not a paradox. It's the result of a working day that consists increasingly of concentrated thinking.
🐛 The problem: AI doesn't reduce work, it intensifies it
AI tools take over routine work. That sounds like relief. In practice it often means the opposite: whoever no longer formats, researches and summarises themselves, evaluates, decides and checks all the more.
A study by Ranganathan & Ye, Harvard Business Review, 2026 shows that knowledge workers with AI support complete more tasks per unit of time but report higher cognitive load than before. Volume rises, density rises, and the pauses in which the brain sorts and regenerates disappear.
McKinsey, 2026 confirms it: teams using AI tools intensively report a 35 per cent increase in daily decisions compared to teams without AI. And the Work Trend Index 2025 shows that 68 per cent of knowledge workers struggle to find enough time and energy for focused work.
🦋 What lies behind it: cognitive capitulation
We described the pattern in an earlier article: cognitive capitulation. The moment when the brain simply waves AI outputs through without checking them, because the capacity is gone. Now we see the flipside: whoever genuinely checks and decides pays a price. The day becomes denser, not lighter.
That is no argument against AI. The point is to use it consciously.
🌱 What leaders can do
It is a design problem: how do we shape work so that AI support genuinely relieves rather than merely raising volume?
🦋 Build in buffers, don't optimise them away. When AI makes things faster, the reflex is to fill the freed-up time immediately. That reflex is the enemy. Deliberately protect time without new tasks.
🦋 Decision quality before decision speed. More decisions per day is not a success indicator. The question is: which decisions are genuinely well-considered? AI outputs that go unchecked increase risk, not quality.
🦋 Make cognitive roles explicit. Who in the team checks, who decides, who approves? These roles become more important with AI, not less. When everyone uses AI simultaneously and no one explicitly checks, errors multiply.
🦋 Create space for learning. Writing by hand, formulating ideas yourself, wrestling with a statement: none of these are inefficient detours. As Ernest Hemingway put it: the only way to learn how to write is to write. (Paris Review, 2004). Whoever lets that ability atrophy loses the foundation for sound judgement.
What this looks like in practice
A consultancy with 20 employees introduces AI tools for research and summaries. The first week feels productive. After a month the team reports back: the days are fuller, not lighter.
Rather than withdrawing the tools, the firm focuses on design. Every task explicitly gets a phase without AI: the initial thinking, formulating the right question, the final review. The AI phase sits in the middle. Volume stays high, but the cognitive structure gives the team rhythm again.
Benefits at a glance
🦋 Less exhaustion through deliberate design of AI use rather than reactive adoption
🦋 Better decision quality, because review processes remain explicit
🦋 Preservation of critical thinking skills that don't atrophy through AI delegation
🦋 A team that leads AI as a tool, rather than being led by it
Conclusion
AI makes more possible. Making work lighter is not automatic. That is a leadership task.
👉 How is it for you? If you notice your team becoming more productive but more exhausted, a conversation is worthwhile. Feel free to get in touch.
A concrete example from the healthcare sector: AI in Spitex documentation: time saved, energy lost?