Seven out of ten managed service providers (MSPs) say they’re using agentic AI. But dig into where they’re actually using it, and the MSP agentic AI adoption picture is far less impressive. Only 10% have primarily deployed it in IT service desk or security and compliance operations. The rest are tinkering internally while their customers, in many cases, are already further ahead.
That’s the headline finding from recent data published in “The Autonomy Advantage,” a white paper from Omdia (commissioned by SuperOps) that draws on multiple polls of MSPs and IT decision-makers conducted through the Candefero and Canalys databases in mid-to-late 2025 to better understand MSP agentic AI adoption.
MSP Agentic AI: The gap between experimentation and execution
In my view, the report’s headline claim that agentic AI will be the biggest operating model shift since cloud isn’t really the story here. It’s about where agentic AI is and isn’t being used.
As already mentioned, according to Omdia’s data, 70% of MSPs report using agentic AI internally. This sounds impressive until you dig into where: only 10% have primarily extended agentic AI to their IT service desk or to security and compliance operations. Meanwhile, 30% of MSPs report no use of agentic AI at all.
That’s a significant execution gap. Most MSPs appear to be in a holding pattern: aware of the technology, possibly experimenting with it, but not deploying it in the workflows that would benefit most. The report calls this a divide between “the agentic AI vanguard and the agentic AI delayers,” which is a bit buzzwordy for my taste, and it misses the larger middle group. A large number of organizations are stuck in experimentation without a clear path to production use.
On the end-user side, the picture is slightly more advanced. Omdia found that 13% of IT organizations’ customers are primarily using agentic AI for customer experience, another 13% for productivity, and 9% for business process automation. Only 17% aren’t using it at all. So customers are arguably moving faster than the MSPs serving them, which creates its own set of pressures.
The cost argument is getting harder to ignore
Omdia’s modelling suggests that enterprise deployments of agentic AI for employee support (including IT services) could deliver annual cost savings of $250–1,200 per employee. That’s a wide range, and the report doesn’t break down exactly what drives the variation. Still, even at the lower end, the savings add up quickly across a large workforce.
The cost case is most obvious for MSPs, where Omdia estimates around 40% of costs are spent on labor. The report points to Level 1 ticket resolution as the obvious starting point: password resets, basic software installations, account unlocks, and printer issues. These are the repetitive, manual tasks that eat up technician time without moving the needle, and they’re precisely the kind of work that agentic AI can handle autonomously or semi-autonomously.
For internal IT teams, the main benefit is faster ticket resolution and the ability to offset headcount constraints. Automating components of multi-factor authentication (MFA) setup, email configuration, and access problem resolution means users experience shorter downtime, which directly feeds into satisfaction scores and overall enterprise efficiency.
MSP Agentic AI: Governance is the real barrier, not technology
One of the more interesting findings is that governance and compliance is the top barrier to implementing agentic AI, cited by 47% of respondents. Technical expertise gaps accounted for just 16%, with value realization and adoption at 14%, data and security management at 14%, and business integration challenges at 9%.
Think about that for a moment. The technology itself isn’t the problem. The challenge is that organizations don’t yet have the frameworks in place to manage autonomous systems responsibly. Questions about data access, model transparency, auditability, and when human override should kick in remain unresolved for most. For an industry that has spent decades building governance structures around ITIL processes and service management frameworks, this should feel like familiar territory. But applying those governance instincts to AI systems that can reason and act independently requires new thinking.
The report recommends establishing governance frameworks early rather than treating them as an afterthought, a sensible approach that too few organizations seem to follow.
MSP Agentic AI: The maturity curve
Omdia outlines a three-stage maturity model for agentic AI adoption: assistive, semi-autonomous, and autonomous. In the assistive stage, AI classifies issues, recommends actions, and speeds up human-led workflows. In the semi-autonomous stage, AI executes routine tasks with human approval. In the autonomous stage, systems coordinate actions across IT, security operations, and cloud operations with minimal supervision.
Most organizations sit firmly in the first stage, with some beginning to move into the second. The leap to genuine autonomy, where systems make cross-domain decisions and execute end-to-end workflows independently, remains largely aspirational for most.
What’s more useful in practice is the report’s distinction between MSP and internal IT roadmaps. For MSPs, the starting point is integration: aligning remote monitoring and management (RMM) and professional services automation PSA platforms so that data flows cleanly between monitoring, ticketing, and billing systems. Without that unified data foundation, agentic AI doesn’t have the context it needs to make reliable decisions. For internal IT teams, the starting point is an audit: identifying repetitive tasks, recurring incidents, and cross-departmental dependencies that slow resolution times.
Both paths emphasize starting small and building trust. Pilot programs targeting high-volume workflows such as alert management, ticket triage, and patch automation deliver quick wins that build organizational confidence and justify further investment.
AI maturity and MSP valuations
In my opinion, one finding that deserves more attention than the report gives it is the connection between AI operational maturity and MSP business valuations. The paper notes that private equity investors and acquirers are beginning to factor AI maturity into their assessments. For MSPs considering exit strategies or seeking investment, this means delaying agentic AI adoption isn’t just an operational risk; it’s a strategic one as well (and potentially a financial one, too).
The report argues that these differences compound over time, creating a widening gap between organizations that are operationalizing agentic workflows and those that are merely experimenting with them. Early adopters gain structural cost advantages and faster service delivery, while lagging organizations find it increasingly difficult to catch up.
Reading between the lines
It’s worth noting that this report was commissioned by SuperOps, who naturally have a commercial interest in promoting the adoption of AI-enabled IT management tools. The poll data comes from the Candefero and Canalys databases, with each poll surveying between 300 and 400 decision-makers, which gives the findings a reasonable level of credibility. However, as you’d expect from a vendor-commissioned piece, the emphasis is firmly on why you should be adopting now rather than waiting.
That said, the underlying data points are useful for anyone trying to understand where the market is heading. The execution gap between internal experimentation and production deployment is real. The governance barrier is real. And the competitive pressure on organizations that delay is, if anything, likely to intensify as the technology matures.
The full report includes detailed action plans for both MSPs and internal IT teams, along with the complete polling data and Omdia’s formal recommendations. You can download it here.
Whether you’re already piloting agentic AI in your service desk operations or still weighing up the options, the window for early-mover advantage is narrowing. The question isn’t whether agentic AI will reshape IT service delivery and support, but whether you’ll be ahead of the curve or playing catch-up.
Further Reading
Sophie Danby
Sophie is a freelance ITSM marketing consultant, helping ITSM solution vendors to develop and implement effective marketing strategies.
She covers both traditional areas of marketing (such as advertising, trade shows, and events) and digital marketing (such as video, social media, and email marketing). She is also a trained editor.
