SolarWinds says AI in IT service management hides costs

SolarWinds has published research suggesting most IT teams underestimate the cost of adopting artificial intelligence in IT service management. Only 7% of respondents said adoption costs matched their plans.

The survey of more than 800 IT professionals worldwide also found that 84% said AI had met or exceeded return on investment expectations, highlighting a gap between perceived value and the operational burden of maintaining the tools.

That tension runs through much of the report. While respondents reported time savings across several service management tasks, many also said the technology had added new work rather than removed existing work.

More than half, 52%, said their overall workload had increased since adopting AI in IT service management. Another 83% said they now spend three or more hours each week keeping AI systems running reliably.

The findings suggest that for many IT departments, the economics of AI in service management are being shaped less by software licences than by staffing, data preparation and ongoing maintenance. Those costs appear to be recurring rather than one-off implementation charges.

Hidden costs

Among the biggest unexpected expenses, 48% of respondents cited staff training, 47% pointed to data quality and cleanup, and 45% highlighted ongoing tuning and maintenance. The figures suggest organisations are encountering practical problems after deployment, even when headline productivity measures appear positive.

IT teams said the time saved by AI is often being redirected into new operational duties. Respondents reported average weekly savings of 3.2 hours on detecting and flagging issues, 3.0 hours on handling end-user requests and 2.9 hours on ticket triage.

But those gains were offset by extra work managing the systems. The survey found that 48% spend time managing and maintaining AI tools and integrations, 47% review and validate AI-generated outputs, and 37% train and fine-tune models.

This helps explain why many teams report a heavier workload despite measurable efficiency gains. In effect, AI appears to be changing the shape of service management work rather than reducing the total volume of work required.

Reactive use

The research also suggests many organisations are still using AI reactively. When asked where AI had the greatest impact across the incident lifecycle, 31% of respondents chose identifying issues before they affect users and 23% chose prioritising and routing issues.

Only 19% said AI had its greatest impact in preventing issues before they occur. That points to more limited use of the technology, with many IT teams applying it to incidents that have already emerged rather than to prevention.

SolarWinds linked that pattern to broader questions about AI maturity, saying the tools may be in place but many organisations still lack the infrastructure, data foundation and internal processes needed to shift AI use upstream.

Budget trends indicate that companies are still willing to spend more. According to the survey, 85% said their AI budget within IT service management had increased year on year, while 36% said it had increased significantly.

That rise in spending is happening even as teams report strain on staff and processes. The data suggests many organisations still see enough value in AI to expand investment, but are grappling with the practical demands of making it work in production settings.

Workplace pressure

The report also points to growing pressure on employees to turn AI adoption into measurable gains. Some 66% of respondents said their bonuses and performance reviews were tied directly to AI efficiency gains.

At the same time, 82% said their organisations offer formal AI training and structured change management. That combination of incentives and training suggests businesses are trying to embed AI more deeply in routine service management operations.

Another notable finding concerns how companies assess results. Only 21% of respondents said they measure AI in outcome or experience terms rather than by activity.

According to the survey, teams that measure AI by activity instead of outcomes were 2.4 times more likely to say their workload had increased since adoption. That points to a management issue as much as a technical one, with measurement practices affecting whether AI is seen as reducing effort or simply adding tasks.

Brad McGinity, General Manager of ITSM at SolarWinds, said the sector had reached a turning point. "We're at an inflection point in IT service management. AI adoption is no longer the hard part - the hard part is building the organisational discipline to make AI actually deliver," he said.

He added: "The teams that get this right aren't just running a faster service desk; they're running a fundamentally different operation. At SolarWinds, our job is to make that transition as straightforward as possible - giving customers the platform, the data foundation, and the governance they need to move from AI activity to real AI payoff."