Leaders are burning out on AI without being able to name the cause. The reason is that AI has made it possible to produce far more of everything, including things that produce nothing, and a leader cannot tell which kind of request they are looking at until after they have answered it. Either way it costs them the same.
Two kinds of burnout, and only one of them is worth having
There is a positive version. People and AI together genuinely produce more and better work: more profitable, fewer wasted resources, real output. It is tiring in the way that a good quarter is tiring.
Then there is the negative version. The organisation generates more questions, more requests and more demands than it can absorb, none of which lead anywhere, and all of it feels productive because the volume is real. Everyone is busy. Very little changes.
The trap is in the middle. A request is often 35 per cent useful and 65 per cent noise, and the leader cannot separate the two until they have engaged with it. The evaluation costs the same as the answer. So the fatigue arrives regardless of what they decide.
Your KPIs are now the problem
If people are measured on production, AI lets them hit every target while producing more of nothing. Nav’s 2026 employer survey found that 53 per cent of Norwegian businesses now use AI for at least one work task, so this is no longer a problem confined to early adopters.
That is not a workforce failing. It is an incentive working exactly as designed, in a world where the cost of producing something has collapsed. Output is no longer scarce, so measuring it no longer distinguishes anyone.
This is the moment to rewrite KPIs around outcome rather than output, because pass-fail measures quietly destroy the value AI produces on the way past. It is unglamorous work and it sits with HR and management rather than with technology, which is precisely why it keeps being postponed.
Underneath it is a cultural problem. Most people come from a world where as long as they produced, they were doing the job. AI finishes today’s work so quickly that there is no longer an excuse not to be looking at next week and next quarter. People trained to produce for today do not always understand this, and it costs time, resources and client credibility. A company using AI seriously needs people with vision, not people using AI to cover their targets.
The new pressure nobody had before: your client is watching you with AI
This one is genuinely new and I have not seen it written about.
Suppose you built your client a monitoring system. Your client has now pointed AI at your delivery of that system, and monitors your monitoring. Every service level, every contractual condition, watched continuously by something that never sleeps, never forgets and never lets a breach pass unnoticed.
This did not exist before, because clients did not have the people or the hours to do it. Now they do not need either. If you go down, or drift outside a condition, they can act on it under the contract immediately.
It shifts the balance of power between supplier and customer, and it introduces a form of continuous scrutiny that leaders have never had to carry. Worth asking, before the next renewal, whether anyone on your side is reading your contracts the way an AI would.
Tool lock-in burns people out too
A smaller but common source. A company is restricted to one vendor’s assistant because of an enterprise agreement, while employees know perfectly well that other tools do more of what their work actually needs. That is one of the quieter reasons people start using AI where you cannot see them.
Forcing an inferior tool exhausts the employee, exhausts the leader, and produces worse output. Worse output generates more bad requests, and the whole organisation ends up busy with something that should never have been asked in the first place. The licence saved money on a line item and spent it three times over somewhere less visible.
Somebody has to hold this
Burnout here is not confined to the executive floor. Culture, governance, HR policy and management KPIs all have to be reworked, company by company, because no two organisations measure people the same way.
That is a large piece of work and it does not belong to IT. Without it, AI never really lands, and everyone gets tired proving it.
The question worth putting to your leadership team this month is simple. What are we measuring, and would our people still hit those numbers if the work behind them had to matter?
Questions this article answers
Why is AI making leaders more tired rather than less?
Because it multiplies requests and questions faster than an organisation can absorb them, and a leader must engage with each one to discover whether it was worth engaging with.
What is the difference between positive and negative AI burnout?
Positive burnout comes from genuinely producing more of value. Negative burnout comes from producing more volume that leads nowhere while feeling productive.
How should KPIs change now that people use AI?
They should measure outcome rather than output. When producing something costs almost nothing, output targets no longer distinguish good work from noise.
How are clients using AI to monitor suppliers?
By pointing AI at supplier delivery and contractual conditions, so performance is watched continuously rather than reviewed periodically. This was previously impossible for reasons of cost and staffing.
Can restricting employees to one AI tool cause harm?
Yes. Forcing an unsuitable tool produces worse work and more unnecessary requests, and the cost of that usually exceeds whatever the licensing arrangement saved.
If you are leading your organisation through this, this is the kind of work I do with a limited number of senior leaders each quarter. → https://anglero.com/the-reset/
Strategic AI Advisor, MerkabaPhi AS, Oslo. 450+ keynotes across 30+ countries.
Listen to the Clarity at the Top podcast on Spotify.
Enquiries: anglero.com