There's been no shortage of AI commentary aimed at MSPs over the past year, but much of it tends to fall into one of two categories. It's either very high level and abstract, focusing on broad trends without much practical grounding, or it becomes overly focused on specific tools, losing sight of the wider commercial and operational shift that's taking place.
That's why our CEO, Terry Lewis, recently sat down with Mostyn Thomas, Senior Director of Security for EMEA in the Academy Department at Pax8, to explore how MSPs are adapting to AI in practice, and what the shift towards the Managed Intelligence Provider model looks like.
Pax8 has played a significant role in shaping this conversation through its Agentic Inflection Point report and MIP Playbook, both of which have pushed the industry beyond high-level AI discussions and into more practical questions around monetisation, delivery, and value.
This conversation adds a clearer view of how those ideas are translating into real-world MSP operations. Feel free to kick back and watch the video, or read below for a summary on some of the main themes covered throughout.
đź”— The full video can be found on YouTube via this link: https://youtu.be/KT6jxxfR_Rg?si=A_oaMJ1V9DcQ3hja
What came through clearly in our conversation with Mostyn is that this change isn't about whether MSPs should be using AI. Most already are, in some capacity, whether through automation, internal tooling, or early-stage customer use cases. The more pertinent question is how AI is reshaping the MSP model itself, and what that means for how providers deliver and capture value.
This is where the idea of the Managed Intelligence Provider has gained a lot of traction. It's an extension of the MSP model (rather than a replacement), that places greater emphasis on managing intelligent systems, orchestrating automation, and delivering outcomes that are tied more closely to business performance . What's often missing from the broader conversation, however, is how that move is actually playing out in practice.
“If [MSPs] don’t get on this journey and get at least somewhere, they will be left behind or taken over by someone who is doing that.” - Mostyn Thomas, Pax8
A consistent theme throughout our conversation with Mostyn was the gap between awareness and execution. There's very little ambiguity at this point about where the market is heading - AI is already reshaping how services are delivered, how value is defined, and how customers engage with technology providers.
But so far, adoption remains uneven. As Mostyn points out, most MSPs are aware that they need to be on this journey, and many of the more forward-leaning MSPs are already building automation into their operations using platforms such as Claude Code, n8n, and rewst, albeit not as fully productised services yet.
But the practical reality is that running an MSP leaves little room to step back and rethink the model while continuing to deliver day-to-day services.
“Nearly all MSPs are aware that it's happening and they need to be on that journey. But reality bites... Most MSPs haven't got a spare five minutes in the day.” - Mostyn Thomas, Pax8
At the same time, there's a growing recognition that this shift can't be deferred indefinitely. Previous transitions in the industry, such as cloud, telecoms, and managed print followed a similar pattern. But AI differs in one key respect: It's being driven as much by end users as it is by vendors. Customers are already adopting AI tools in their own workflows, which accelerates expectations and compresses the timeline for providers to respond. As such, while it may still be early, the "ubiquitous" and user-driven nature of AI makes it increasingly difficult for providers to remain passive.
"It's about outcome." - Mostyn Thomas, Pax8
The general conversation around monetisation with AI is, in many cases, unresolved and uncertain. As Mostyn and Terry discussed, moving from service-based billing to outcome-based value is often presented as a natural evolution, but in practice it introduces a level of complexity that MSPs have not had to deal with before. Pax8’s Managed Intelligence Provider (MIP) Playbook is particularly relevant here as it moves beyond theory and starts to outline how MSPs might package, price, and deliver these new forms of value.
Under a traditional model, services are delivered within clearly defined boundaries, and are typically tied to infrastructure, licenses, or time. If something falls outside of scope, or if an outcome isn't achieved, there's a degree of separation between the provider and the result. Moving towards outcome-based models introduces a more direct relationship between delivery and value.
It also requires a different set of capabilities internally. Pricing, sales, and service delivery all need to adapt to a model where value is less tangible and more closely tied to business performance.
The questions MSPs are asking reflect that uncertainty:
How do you package something like an AI agent into a repeatable service?
How do you ensure recurring revenue when the underlying work may be project-based or variable?
And how do you communicate that value in a way that customers understand and trust?
Nonetheless, there is a clear precedent. Similar concerns were raised during the shift to cloud and, later on, during the early adoption of platforms such as AWS and Azure, where the immediate reaction was uncertainty around how value would be maintained. In both cases, MSPs found ways to package and monetise those services effectively over time.
The shift towards AI introduces both risk and opportunity for smaller MSPs:
“The challenge for smaller MSPs is that customers may go direct, or a larger MSP may come in and take that opportunity instead.” - Mostyn Thomas Pax8
On one hand, there's a credible concern that customers may bypass smaller providers by going directly to vendors that have already developed AI capabilities.
Though on the other hand, the same forces that create that risk also reduce traditional barriers to entry and open up new opportunities. The ability to prototype, test, and build services has become significantly more accessible. What previously required substantial investment can now be explored internally, often with minimal upfront cost, moving the competitive dynamic away from scale and more towards execution.
Automation plays a central role in this; as more work is handled by systems, workflows, and agents, the output of a small team can increase, disproportionately so. Mostyn's example that a four-person MSP can begin to operate with the output of a much larger organisation is very much a reflection of the broader shift towards what could be described as a 'digital workforce'.
“It’s really around intelligence and data, and how you can make that work for you, rather than how much money you can invest.” - Mostyn Thomas, Pax8
"We're seeing more MSPs using things like Microsoft Sentinel to sort of build their own SOC and SIEM capabilities." - Terry Lewis, RoboShadow
Another important theme in the conversation was the gradual democratisation of capabilities that were previously limited to larger providers. The example of SOC and SIEM services illustrates this clearly. Tools such as Microsoft Sentinel, combined with Defender, Intune, and increasingly Copilot-driven automation, are making it possible for MSPs of all sizes to deliver capabilities that in the past were reserved for larger providers, requiring significant internal resource.
Microsoft’s introduction of bundled offerings such as E7, which bring together security, compliance, identity, and AI-driven agent governance into a single platform, points towards a model where capability is increasingly packaged (while still requiring orchestration and oversight).
This all indicates a broader trend in how technology is evolving. The role of the MSP is transforming from building and maintaining systems entirely in-house, to orchestrating a combination of platforms, automation, and processes. In that context, developments such as Microsoft’s integrated security and AI stack, including agent management, make advanced capability more accessible, but don’t remove the need for coordination to ensure it’s applied effectively.
Following the same example used earlier, the move to cloud actually increased the amount of work required to manage environments, as more tools, more configuration, and more dependencies created additional layers of responsibility. AI is likely to follow a similar pattern.
“One thing MSPs can definitely do is work it out. How to make a living and how to add value." - Mostyn Thomas, Pax8
The conversation around internal tooling and “vibe coding” also highlights a broader shift towards experimentation becoming a standard part of MSP operations. The ability to test ideas, build proof-of-concepts, and iterate quickly has become significantly more accessible, allowing providers to explore new approaches without committing substantial resources upfront.
“What would have cost a hundred or two hundred thousand, you can now test yourself, decide what works, and then build something around it.” - Mostyn Thomas, Pax8
This changes how services are developed - rather than relying entirely on external development or waiting for fully formed solutions, MSPs can take a more incremental approach, validating concepts before scaling them. It introduces a more iterative way of working, where ongoing experimentation is part of how the business evolves.
Of course, there are limits to this approach. Moving from experimentation to production requires structure, governance, and, in many cases, external expertise. This reinforces the importance of partnerships, particularly for smaller MSPs - As Mostyn said, “you don’t have to have every skill in-house - you can partner with people who specialise in those areas and still deliver the service".
As AI becomes more integrated into business processes, the role of the MSP extends beyond managing systems to influencing how those systems are used. This requires a deeper understanding of the customer’s operations and a shift towards more strategic engagement.
In practical terms, this means moving closer to board-level conversations, where value is discussed in terms of outcomes rather than inputs. Providers that are able to engage at this level tend to become more embedded within the organisation, and it also provides a potential route away from commoditisation for MSPs who feel they've become intermediaries for larger vendors.
“Data governance and supply chain risk are already front and centre right now.” - Mostyn Thomas, Pax8
This is already reflected in the growing number of frameworks emerging around AI, including ISO 42001 and the EU AI Act, alongside existing standards such as ISO 27001, NIS2, and Cyber Essentials, which MSPs are already helping customers navigate. While this introduces additional complexity for end users, it also creates a natural extension of existing MSP services.
As organisations rely more heavily on automated systems and agents, they need to be able to demonstrate control, accountability, and trust. This introduces new requirements around monitoring, policy, and compliance, which align closely with existing MSP capabilities.
The opportunity here is that governance work tends to operate at a higher level within the organisation, involving senior stakeholders and strategic decision-making. It also creates ongoing demand, as compliance is not a one-time activity but a continuous process; supporting customers in navigating these requirements therefore positions the MSP as a strategic partner.
Terry put it best: “there’s clearly something different about Pax8, people don’t talk about it like a normal vendor.”
Mostyn and Terry spoke about how Pax8's masterclasses help shape how MSPs manage cybersecurity and the broader AI transition. They aren't built around a traditional vendor-led / training format, but are deliberately designed with what Mostyn described as a "partner lens", focusing on the realities MSPs face rather than positioning specific products.
The content covers a wide range of topics, from theory and the history of hacking through to risk, board-level communication, product strategy, and partnerships. Partners are encouraged to share their own experiences, including what hasn't worked - which is often where the most useful insight comes from!
“I’ll probably learn as much from the room as they do, because the partners share what’s worked and what hasn’t.” - Mostyn Thomas, Pax8
This approach has extended into other areas including data and AI, as well as sales, which is telling as it reflects a growing recognition of the commercial challenges. Teaching technical teams how to communicate and sell outcome-based services is becoming increasingly important as the market shifts.
This model ties into Pax8's broader position within the channel. The emphasis on community, supported by a team with significant MSP experience, creates a context in which partners aren't simply consuming information, but actively shaping it. In a market that's still evolving, this kind of shared experience is extremely valuable.
The shift from MSP to MIP is an ongoing process shaped by a combination of technological change, customer behaviour, and commercial pressure. What came through clearly in this conversation is that the direction we're headed in is already well understood, but the challenge lies in how that translates into day-to-day operations.
For most providers, this means gradually adapting how value is defined, delivered, and communicated, whether that’s through experimentation, changes in pricing models, or moving closer to the business outcomes their customers care about. But these aren’t overnight changes, and they won’t look the same for every MSP.
The providers that navigate this effectively are likely to be those that treat this as an extension of what they already do, building on existing capabilities while evolving how they apply them. The opportunity is in expanding the traditional MSP in a way that brings them closer to the core of the organisations they support.
If you’re already exploring this, or even just starting to think about where it fits, you’re not alone. It’s something the entire market is working through.
Thank you for your ongoing support and feedback, and if you have any questions, please don’t hesitate to reach out to us at hello@roboshadow.com