For years, we have been pitched the same promise about artificial intelligence. Once AI takes on routine tasks, people will be free to concentrate on more strategic work. It is central to the business case for automation, yet until recently we have had to accept it at face value. Most deployments were too new to show what happened to human work over time.

IT service management now lets us test that assumption. Respondents to the 2026 SolarWinds State of ITSM Report had used AI in their ITSM environments for an average of about 16 months, and nearly two-thirds had at least a year of experience. These are live settings where AI helps teams detect issues, triage tickets and respond to requests. The data shows what happens once automation becomes everyday work.

The first finding to jump out is that workloads have not fallen for most teams. Fifty-two per cent say their overall workload has increased since adopting AI, while another 19 per cent say it has stayed about the same. On its own, that sounds like the headline, and sceptics may treat it as proof of failure. Dig deeper, however, and it could support the original promise. If AI is moving people towards more strategic work, their workload would not necessarily shrink. It would change shape.

Busy is busy, whether someone is triaging routine tickets or tackling a problem that genuinely stretches them. The flat-or-rising workload figure therefore does not answer the question that matters. It confirms that something moved. What deserves examination is where it went.

The direction of the AI Shift

Here the evidence becomes less reassuring. Work has shifted, but much of it appears to have moved laterally rather than upwards. Almost half of respondents say managing AI tools and integrations has added to their workload, with a similar proportion pointing to reviewing and validating AI-generated outputs. Others are training models, handling failures and governing decisions.

Reliable automation requires oversight, and governance or validation carries value. But this differs from the progression many people imagined. Instead of moving from resolving incidents to preventing them, someone may move from completing a task manually to checking whether AI completed it correctly. The activity has changed without travelling far up the value chain.

There is room for disappointment in that finding. But it is not a verdict on AI, because AI itself never determined the direction of the shift. That emerges from choices about its operating environment, what it is asked to improve and how success is judged. For many teams, those choices appear to be happening by default, rather than by design.

Data determines how much work survives

Data quality is the mechanical fork in the road. The report identifies insufficient data quality as the leading reason AI underdelivers. If the information feeding a system is incomplete or inconsistent, AI may produce an answer faster without producing one that can be trusted. People then spend the saved time checking outputs and correcting mistakes.

The burden has not disappeared. It has moved from manual execution to manual correction, perhaps the clearest example of a lateral shift in the report. Stronger data reduces the checking and repair required around each automated task. More of the saved hour then survives to become something else.

Measurement decides what counts as progress

What happens to that capacity depends partly on what an organisation chooses to notice. Only 21 per cent of respondents describe their measurement as focused on outcomes or experience. Most concentrate on activity or productivity, such as ticket volume and resolution time. Organisations using activity-oriented measurement are 2.4 times more likely to report increased workloads than those taking an outcome-oriented approach.

This finding shows an association rather than proving that activity measures cause workloads to rise. It still explains why measurement matters. A team judged by how many tickets it resolves has a clear incentive to use AI to resolve more tickets. That may increase capacity and improve service performance, but it can also keep people occupied with a growing volume of the same work.

Outcome measures create a different test. Tracking prevented incidents, service quality, employee experience or business disruption requires leaders to consider whether AI changed the work, rather than merely accelerated it. The destination of human effort becomes part of the definition of success. The upward shift must then be identified and designed rather than assumed.

Direction begins with where AI is applied

The choice of starting point compounds the effects of data and measurement. The report finds the clearest returns where workflows are structured, frequent and measurable. Teams can recognise good performance and identify errors without creating a large parallel process around the technology.

AI added to fragmented tools or poorly understood workflows creates integration points and exceptions for people to manage. These consume capacity that automation was meant to create. Pointing AI at a clear process cannot guarantee higher-value human work, but it reduces the drag that prevents work from moving upwards.

The upward shift must be designed

AI adoption will continue, as it should. ITSM shows that it can save time, improve performance and deliver a return. It also shows why automation alone cannot fulfil the wider promise made on its behalf. AI can take on a task, but it cannot decide what work deserves the human time released from it.

That decision remains with us. Data quality determines how much saved capacity survives. Measurement influences where it goes. The workflows chosen for automation determine how much upkeep is created along the way. If these choices are left to existing habits and short-term pressures, work will still shift, but it is likely to travel sideways.

The opportunity is to make the direction deliberate. The redistribution of work is no longer hypothetical, and we are beginning to quantify it. Whether it moves laterally or upwards now depends on treating higher-value human work as an outcome that leaders must actively create.

By Abdul Rehman Tariq Butt, Regional Director – Middle East at SolarWinds

Related post

View all