What AI Is Actually Changing About Managed Services And What It Isn’t

For midmarket IT leaders, the real question isn’t whether their MSPs are using AI—but whether that use is changing outcomes that they can see, measure and trust.

MSPs are adopting AI at a fast pace, and 85 percent of midmarket companies now depend on them for critical IT operations, according to an Azafran Partners report.

But how much of that AI adoption translates into a meaningfully different experience for midmarket customers? Are core elements of service delivery actually changing, or is much of the impact confined to internal efficiency gains within MSP operations?

MES Computing spoke with industry leaders to examine where AI is materially changing managed services today and where the traditional MSP services remain largely unchanged.

Where MSP Delivery Is Shifting

Several industry leaders told MES Computing that AI is making its most tangible impact on the front line of service delivery, where MSPs handle incoming issues and decide how to respond.

Their response is supported by an investment report that shows companies implementing agent orchestration systems are seeing 40 percent to 90 percent reductions in resolution time.

Jay Mellon, co-founder and CEO of MSP AtNetPlus, shared with MES Computing how those gains automate front-line operations and allow more time for their engineers to take up higher challenges. “The biggest impact is that these capabilities free our engineers from repetitive administrative work so they can focus on solving complex problems and advising clients,” Mellon said.

Interim managing director and executive chair at Reliance Cyber, Jenny Kalenderidis, said the shift can also be felt in how MSPs offer cybersecurity services. One of the key changes AI has brought into managed services, she noted, is that security teams, for instance, can now take what it learns from one customer incident and improve detection rules across the entire customer base almost immediately.

“An attack thwarted in one environment immediately strengthens the defenses for all, drastically reducing the time it takes to protect all customers from emerging threats,” she told MES Computing.

What Hasn’t Changed And Why It Still Matters

While AI is clearly changing how MSPs operate internally, some leaders MES Computing spoke with said the parts of managed services that matter most to customers have not changed.

“MSPs are often deeply embedded in their clients’ operations, and executives rely on them to translate technology risks and opportunities into business decisions," Mellon said. “That human layer isn’t something AI replaces.”

Strategic guidance on cybersecurity risk and business continuity planning, he added, still requires the kind of context and judgment that automation cannot deliver.

Paul Cullimore, solutions director at IT security service firm Quorum Cyber, said the company’s customers have been direct about this and are not comfortable with AI making automated changes to their environment, at least not yet. Cullimore told MES Computing that “this will change in the near future because responses to attacks need to accelerate to machine speed, but experienced humans will still be monitoring these AI agents and platforms to ensure true mitigation is taking place.”

Midmarket expectations around cost also reinforce why the human layer still matters despite the changes AI may bring to managed services.

A recent report on how MSPs will evolve with AI found that most organizations are not willing to pay more simply because AI is layered on top of MSP offerings. The report noted that while IT leaders expect MSPs to use AI to improve efficiency while keeping costs under control, if they experience the same service with less human engagement, they will question its value, regardless of what is happening behind the scenes.

The path forward, according to Kalenderidis from Reliance Cyber, is to establish a healthy balance between AI usage and human oversight. In her words, “AI, for instance, can clear hundreds or thousands of daily alerts, and that is genuinely valuable, but thereafter, it’s down to human interpretation to act on this intelligence.”

Key Takeaways For Midmarket IT Leaders

If MSPs are increasingly automating their delivery with AI, midmarket IT leaders need to know how to tell whether that adoption is improving what they receive.

Here is a quick rundown of what IT leaders should scrutinize when their MSP waves the AI flag in client conversations:

1. Ask what has actually changed in your service since AI was introduced.

Head of big data engineering at Innowise, Philip Tikhanovich, said it comes down to proof. “The primary focus should be on measurable improvements and accountability, not simply adopting an AI capability,” he told MES Computing. IT leaders should be asking their provider whether resolution time, incident volume or service availability have measurably improved, and they should expect that data be shared consistently to back things up.

2. Is AI improving your experience or your MSP’s margins?

IT leaders should be asking whether the efficiency AI creates is being passed through to them or absorbed internally by the MSP. “Organizations need to have visibility of how AI models are using their data, including whether or not that data leaves the organization or contributes to a common training set for the provider,” Tikhanovich told MES Computing.

3. Scrutinize AI like any cloud service.

Midmarket IT leaders who would not on-board a cloud vendor without reviewing data residency and compliance should apply that same standard to the AI tools their MSP is running inside their environment. Cullimore said data leakage, privacy and where an AI service is geographically hosted all need the same diligence that any cloud-based service would receive.

4. Don’t let AI governance fall off the agenda after on-boarding.

A TSIA reportcaptures the concept of “AI Operational Debt,” which frames the maintenance burden that AI systems create through model drift, data integrity issues and shifting regulatory requirements. Over time, this debt is expected to accumulate and, if left unmanaged, could erode the return on investment that AI was supposed to deliver.