AI agents are no longer theoretical in IT service management (ITSM). They are actively moving from experimentation into operational reality. But for all the buzz and potential, adoption is not frictionless.

In modern, omni-channel service environments, where the service desk increasingly sits at the heart of both employee and customer experience, that friction is as much about trust and operating models as it is about technology.

Research shows overwhelming openness among UK IT professionals (93%) toward integrating AI agents into ITSM workflows. Simultaneously, a similarly strong majority (92%) express concerns about implementation, governance, and control. This tension captures the duality of where the IT sector stands today: optimistic about the potential of AI, but cautious and pragmatic about its risks.

Few departments are as technically prepared to lead the AI charge as IT. Teams that have long embraced automation, scripting, and data-led operations are now being asked to take a bold next step: integrating autonomous or semi-autonomous agents into service delivery.

For many organisations, that means rethinking the service desk as a strategic, experience‑led function rather than a back-office ticket factory and asking how AI agents can help it deliver faster, more personalised, ‘always‑on’ support. Doing so, however, means tough questions and decisions about trust, governance, team structure, and long-term transformation.

The trust issue is real – and rational 

One of the clearest signals from our research is that caution is not rooted in ignorance, but in expertise. Only 32% of professionals were comfortable with AI agents running service operations independently. The majority want a human in the loop, with 49% preferring AI agents to act only with human approval, and 18% seeing their ideal role as purely advisory.

Trust, in other words, is the dividing line. Not because IT teams don’t understand the technology, but because they understand it all too well. From concerns about model reliability to the lack of justifications in certain AI decisions, these are professionals who know exactly where the cracks might appear.

Interestingly, the data also suggests that trust grows with exposure. Professionals at organizations that had already implemented AI features into ITSM workflows were notably more confident in allowing AI agents greater autonomy. Familiarity appears to breed not contempt, but confidence.

This suggests that AI adoption is likely to follow a phased, iterative path, one that gradually hands over responsibility to AI agents as their performance is validated, rather than relying on wholesale transformation overnight. In practical terms, that often means starting with assistive use cases on the service desk (triage, routing, and knowledge suggestions) before moving to more autonomous resolution flows.

Use cases: where AI agents are already delivering value 

While some IT leaders are still in exploratory mode, others are putting AI agents to work in very specific areas. Our research found that process mining and workflow generation top the list of current AI use cases, cited by 50% of respondents. Other popular applications include generating post-incident documentation and handling more complex workflows — all tasks that require structured data processing, consistent execution, and speed.

These use cases are not about flashy innovation for its own sake. They address core ITSM pain points: slow ticket resolution, limited technician bandwidth and inconsistent follow-up processes. By accelerating these repetitive, rule-based tasks, AI agents allow human technicians to focus on the kinds of high-impact work that drive real organisational change. And when embedded directly into the service desk, AI agents can power intelligent ticket handling across channels, recommend next-best actions to technicians, and surface the right knowledge in the moment — the foundations of a consumer-grade support experience for both employees and customers.

Self-service is another area seeing rapid transformation. AI-enabled virtual agents and chatbots are now able to understand intent, trigger automated workflows, and resolve common issues end-to-end. The result is a service-desk that extends beyond the reach of traditional office hours while maintaining consistency and control.

A changing workforce, not a disappearing one 

Concerns about workforce displacement persist at board level, with 59% of senior executives fearing AI will displace human workers. However, sentiment within IT teams paints a more nuanced picture. 62% of professionals believe that AI agents will change their organisation’s hiring plans, but not necessarily create job cuts. Rather, the roles themselves are evolving.

Instead of eliminating technician positions, AI is reframing them. For example, 38% of respondents expect technicians to supervise AI systems, and 30% see a shift toward more complex, strategic IT tasks. This supports the view that AI agents will not replace human professionals but instead augment them, pushing routine work to machines, while elevating the human contribution.

The distinction is important. AI agents aren’t ‘replacing the workforce’; they’re transforming workloads. By taking the manual strain out of ticket categorisation, incident response, and reporting, AI allows humans to focus on service innovation, cross-team collaboration, and systems optimisation. But that’s only possible when IT teams are reskilled, and governance processes are updated to support this hybrid model.

On the front line, the service desk becomes a proving ground for new capabilities: managing AI‑driven workflows, curating training data, tuning virtual agents, and designing support journeys that blend self-service, chat, and human expertise.

Sector-specific attitudes: no one-size-fits-all approach 

One of the more nuanced findings in our research was the variation in AI attitudes across industries. Legal professionals, for instance, showed higher levels of caution, likely due to the high bar for compliance and risk aversion. In contrast, the healthcare and telecoms sectors were among the more enthusiastic adopters, driven perhaps by their need for speed, consistency, and operational scale.

There was also a notable public vs. private divide. Private sector organisations appear more open to embedding AI agents across ITSM workflows, while public sector respondents were more likely to cite data privacy and governance as major roadblocks. These differences matter. They suggest that no universal AI deployment blueprint exists – each organisation and industry must assess where AI agents align with its business goals, operational bottlenecks, and risk appetite. A legal firm and a telecoms provider will likely adopt AI agents very differently, and that’s OK.

What is common, however, is the direction of travel: the service function is moving away from pure SLA compliance and ticket volumes and towards outcomes, experience, and business impact. Across sectors, leaders are asking how AI-enhanced service desks can improve digital employee experience, support omnichannel engagement, and deliver faster, more intuitive support journeys.

Governance, complexity, and the road ahead 

Of course, real transformation won’t happen unless key concerns are addressed. AI governance — including clarity on decision-making authority, auditability, and data use — was the top concern in our research, cited by 45% of respondents. This was closely followed by doubts about the reliability of AI agents (39%) and the complexity of implementing them (34%). These barriers are significant, but not insurmountable. What they reflect is a maturing view of AI in ITSM. This is not a shiny toy, but a fundamental shift in how service operations are run – and that demands strong frameworks, not shortcuts.

To achieve effective AI adoption, three priorities stand out:

● Clear oversight structures – to ensure AI agents operate within well-defined policy boundaries.

● Reskilling strategies – to equip technicians for supervisory, optimisation, and AI lifecycle roles.

● Transparent operations – to help end-users and stakeholders trust that AI is acting responsibly.

When these foundations are in place, the benefits can be profound: faster resolution times, reduced technician burnout, better allocation of strategic IT talent, and higher satisfaction for both users and teams. Crucially, they also prepare the organization to extend AI beyond IT into wider customer service operations, using the service desk as a model for how to blend human and AI support responsibly.

AI adoption is not a leap of faith – it’s a managed transition 

AI agents are already reshaping ITSM, and this transformation is only accelerating. But the narrative doesn’t need to be one of disruption. It can be one of evolution – of better collaboration between humans and technology, smarter service delivery, and more resilient IT operations. In many organisations, the service desk is where this evolution is most visible: from static portals and phone queues to proactive, AI‑augmented, omni‑channel support hubs that feel much closer to modern customer experience platforms.

Recommended reading

• Report: Only 4% of Firms Turn AI Investment into Scalable Value
• AI Bias Locking Women Out of Tech Growth
• UK Business Leaders Struggling to Cope With AI and Cyber Threats

The task ahead for IT leaders is not simply to ‘implement AI.’ It’s to understand what kind of problems AI agents are best suited to solve, build governance that ensures reliability and compliance, and align team structures to make the most of both human and artificial intelligence. That means designing service desks that are data‑driven, experience‑centric, and ready to orchestrate a mix of chatbots, virtual agents, human experts, and back‑end automation.

In doing so, organisations won’t just future-proof their ITSM strategies – they’ll lay the groundwork for a more agile, efficient, and intelligent enterprise. And as expectations for frictionless, personalised support continue to rise, the AI‑enabled service desk will sit at the centre of how organisations deliver value to both employees and customers.