Everyone is talking about MSPs bringing AI to clients. But what does that actually mean?
A client may say they want to “use AI,” but that can mean anything from personal productivity and automation to governance, workflow improvement, or help choosing the right tool. If the conversation starts too broadly, it can turn into a service the client was never ready to buy.
This was a major theme in a recent episode of Above the Stack, where Empath Co-founder Wes Spencer and Cloud Capsule CEO Nick Ross were joined by Clear Guidance Partners Founder Dustin Bolander. Dustin has been actively helping clients use AI in real business workflows, and one of his clearest points was that clients often ask for AI before they know what they actually want from it.
For MSPs, that changes the starting point. AI enablement starts with understanding how the client’s business works, where the work gets stuck, and whether AI is even the right answer.
This article focuses on what MSPs need to understand before they turn AI interest into a client-facing service.
One of the first things Dustin made clear is that the AI conversation with clients does not stay still for very long.
Clear Guidance Partners has been doing more structured AI work with clients for about a year, and Dustin said their offers have already changed roughly every four months. Clients are learning quickly. What felt useful to explain a few months ago may already feel too basic now.
Consequently, AI enablement cannot be treated like a one-time package that never changes. Clients may start by needing help understanding ChatGPT, Claude, or Copilot. A few months later, they may be asking better questions about where AI fits into their workflows, what their employees should be allowed to use, or how to turn small experiments into more reliable ones.
The opportunity is not just to sell an AI tool, but to keep helping clients make sense of what AI can do for the way their business works.
A client saying they want AI is not the same thing as a client knowing what they want AI to do.
Dustin explained that some of the early AI work became difficult when the client came in with interest but little direction. The client knew they wanted to do something with AI, but as the project progressed, it became clear they had not clearly defined the problem they were trying to solve.
AI enablement has to start with better discovery. Are they looking for productivity, better documentation, faster response times, cleaner internal processes, or something else entirely?
Until those questions are answered, “we need AI” is mostly a signal. The client is interested. The service is not yet scoped.
A useful AI conversation starts with, “How does your business work?”
Before an MSP can recommend where AI belongs, they need to understand the client’s business model, the way their team makes money, and the workflows that shape their day-to-day work.
For example, Dustin talked about how different law firms bill in different ways. If a firm bills hourly, using AI to speed up certain drafting work may not automatically create the business value someone assumes it will. In some cases, the better starting point may be administrative work, time capture, intake, or another process that supports the business without creating the wrong incentive.
AI enablement starts with the work the client is already doing and asks where that work could become faster or more consistent.
If an MSP is trying to serve every industry, every department, and every possible AI use case at once, the service can quickly become too broad to explain or deliver consistently.
Dustin pointed to specialization as an important part of making AI work repeatable. Clear Guidance Partners works closely with law firms, so the team does not have to reinvent the conversation each time. They understand the common systems, billing models, document-heavy workflows, client intake processes, and administrative bottlenecks that tend to show up across that vertical.
AI enablement works if you serve one industry. Repeatability needs a narrower frame. With it, the MSP can start building repeatable discovery questions, common use cases, training materials, client examples, and service boundaries.
Sometimes a client asks about AI because they have a real workflow problem. That does not always mean AI is the right fix.
Dustin shared a simple example from a client that needed to create a large volume of similar letters. The client saw it as a potential AI use case, but once the workflow was understood, the simpler answer was clear: mail merge. It saved the team hours without making the project more complex than necessary.
That is an important distinction for MSPs. AI enablement is helping the client understand what they are trying to improve and choose the right solution for that work.
In some cases, that solution may be AI. In others, it may be automation, better templates, cleaner processes, stronger training, or a tool the client already owns but has not been using well.
That makes the MSP’s role more valuable. When the recommendation is based on the workflow rather than a buzzword, the AI conversation becomes more practical and credible.
This article only covers a few of the bigger ideas from the conversation. For the more practical side, watch the full Above the Stack episode on YouTube. They dig deeper into how MSPs can start turning broad client interest into something more useful.
You can also download the Claude 101 resource Dustin prepared as a starting point for client AI education.
And if you want more conversations like this, subscribe to the Empath newsletter for more MSP-focused ideas on client education, service delivery, and the future of AI in the channel.