Most IT service businesses and MSPs are looking at AI completely wrong. They're focused on the tools the added AI feature in their vendor's stack when the real money is somewhere else entirely. I've been working with MSPs daily for years, and I can see a massive gap opening up. The MSPs who figure this out in the next 36 months are going to dominate. The ones who don't will be left behind.
I call it the AI Ascension Model.
Here's what I'm seeing right now: vendors are pushing their AI-powered tools, MSPs are reselling them, and everyone's patting themselves on the back thinking that's their AI strategy. It's not.
The cost of software has dropped to near zero. You can build almost any tool you want with the right AI. The tool is not the opportunity. The opportunity is what you do with AI inside your clients' businesses.
And here's the uncomfortable truth most MSPs are too busy doing IT support, keeping clients compliant, keeping the lights on. They don't have time to develop an actual AI strategy. So it becomes an afterthought. The vendor mentions a feature, there's a flicker of interest, and that becomes "our AI offering."
That's not good enough anymore.
The AI Ascension Model
I've been developing this framework, and I want to share it here in its rough and ready form because I think MSPs need to start thinking about this now.
Step 1: AI Compliance, Governance & Awareness
This is the foot in the door. It's stuff you probably already touch, but haven't packaged properly.
It means making sure there's a whitelist of approved AI tools in the business. There's a process for creating agents and what data they can access. There's a policy around all of it. There are guardrails in place maybe a block list, maybe your client doesn't want staff using certain LLMs. Your web filter can handle a lot of this already.
Then there's the awareness piece training staff on the tools they've got, best practices around data inputs, making sure confidential information isn't being fed into consumer-grade AI, and that there's human verification on outputs.
This is roughly a two-week project. It's the baseline.
Step 2: Data Foundation
This is the unglamorous one. But skip it and everything else falls apart.
You're going into the client's environment their M365 stack, their SaaS apps, their on-site applications and you're finding out where all their data lives. Because here's what I keep seeing: businesses have these big AI project ideas, they go to implement them, and then they hit a wall. The data is organic, scattered, unstructured. The AI systems can't latch onto it.
The data foundation stage is about centralising that data, cleaning it up, and getting it into something like a data warehouse pulling in different sources through APIs or automations into a clean central location the AI actually has access to.
This is also where your managed service lives. Monthly recurring revenue to maintain those data streams, keep the warehouse clean, backed up, always up to date. This is your MRR engine inside this model.
Step 3: Business Process Automation
This is where the hype cycle guys are living right now. And honestly? It's where most MSPs will struggle. But it's also where the biggest opportunity is.
You can take the word "AI" out of this step and just call it what it is: business process automation.
This is not systems engineering. It's not conditional access policies or Azure resources. It's completely different. Your staff aren't trained in it. I'll be honest I'm technical by background, senior systems engineer, and even I had to shift my thinking here.
The opportunity is going into a business, analysing and documenting their processes, and deciding together which ones AI can automate. And yes this does destroy jobs. Let's not pretend otherwise.
But think about it like this. 200 years ago, almost everyone was connected to farming. Then industrialisation happened. Tractors replaced manual labour. Jobs were destroyed. But more jobs were created building the tractors, mining the resources, selling them, maintaining them. The same thing is going to happen here. AI is going to demolish a huge chunk of the white collar job market. But new jobs will be created. And the MSP that helps clients navigate that transition? That's the trusted partner every business is going to need.
Step 4: Transformation
This is the endgame. And it's the hardest to define because every business is unique.
Once you've automated away the processes in Step 3, you've freed up people. Those people and that data become the raw material for something bigger. The goal here is to look at what makes this client's business unique. Their customers, their data, their processes. And then figure out how to use AI to either transform part of their business or create an entirely new revenue stream for them.
That's the north star of the whole model: using AI to generate new revenue for your clients.
Where Does This Leave Us?
I'll be honest I don't have all the answers yet. I put this model through several AIs before I published this and they ripped it apart. Some suggested Step 1 and Step 2 should be combined. Some pointed out that clients will be at different stages of maturity, so you can't just start at Step 1 with everyone some are already past it.
This is a work in progress. But here's what I know: this model, or something that looks like this model, is where the MSP revenue opportunity sits for the next 36 months.
While everyone else panics about AI taking jobs, you can be the person who walks into a business and says: "Yes, AI is going to change everything. And here's exactly how we're going to use it to grow your business."
That's a very different conversation than "here's the AI feature in our monitoring tool."
I'm running seven in-person workshops across the UK in 2026 to develop this framework properly workshopping each stage, stress-testing it against the real world, and building it together with the MSPs I work with. I want to have something closer to a gold standard by mid-2026, but more importantly, I want a community of people who've actually helped build it and are implementing it.
If you're around the world and can't make it in person, I'll be running these virtually too inside the IT Rockstars community over at skool.com/msp. There's a seven day free trial if you want to check it out.