Should Your MSP Sell Apps Built With AI?
Yes, if you sell it as a proof of concept to a client who understands what that means, and you agree up front on who supports it and who owns the next change. Josh Peterson said the first version of a tool he built with an AI assistant showed up about three minutes after he asked for it, and Gary Boyle's warning is that building that fast is the easy part: the first change request is where an MSP that doesn't understand its own build loses. For more context, see how artificial intelligence is transforming MSP software and how to prepare your MSP for the next wave of technological change.
In this BMK Vision Roundtable episode, Josh Peterson, CEO of Bering McKinley, and Gary Boyle, COO and Management Consultant, walk through Gary's levels of AI adoption, the danger zone in the middle, and the choice he believes every MSP now faces.
Should Your MSP Sell Apps Built With AI?
Yes, with conditions, and the conditions are the whole deal. Josh made the case every owner will recognize: you build a tool for yourself in a day, you see clients who need the same thing, and you wonder why the MSP shouldn't build it for them and charge a monthly fee to maintain it. Gary's answer started with a phrase he picked up from a software developer: building for yourself is developing in green pastures. Nobody uses the software yet, no data depends on it, and changing a screen costs nothing.
Once a client relies on it, the first change request arrives, and if you don't really know what you built, you've lost. Josh pushed back with the right client, one who understands the first version will be great and later changes won't come with off-the-shelf guarantees. Gary agreed that is where an MSP can deliver a real proof of concept, as long as the messaging is clear, because the end user will assume it's a finished product.
- Sell a proof of concept, and say so plainly to everyone who will use it.
- Agree before you start on who supports the build and who owns the next change.
- Treat the first change request as the real test of what you sold.
What Is the AI Danger Zone for MSPs?
It's the stage where building is exciting and easy but almost nothing is supportable. Gary lays out a journey. Patient zero has heard of AI and does nothing with it. Level one uses it as a better search engine and a faster writer, and there the MSP's role is mostly advice on tool choice, security, and governance. Level two starts building: interactive tools, scheduled tasks, agents, a daily brief, a dashboard that pulls several business systems together.
That is where MSP project work begins, and it is also where the danger zone begins. Gary's example: the technician who built a client's workflow doesn't show up, another technician has to support something with no documentation, and the CEO is walking into a meeting without the presentation it was supposed to produce. It worked once, maybe for a week. Josh recognized his own early builds, light automations that always felt fragile and rested on instructions he had written months earlier.
- Ask whether a build can be replicated, supported, and secured before you sell it.
- Document every client build so more than one person can maintain it.
- Plan for changes to the AI model underneath a build that alter how it behaves.
Why Does an AI Build Need a Goal and a Plan First?
Because the tools will build in any direction you point them, and they never stop on their own. Gary's view is that an MSP owner who isn't spending serious time in these tools probably won't make it, but time without a goal and a plan just produces output you can't explain. He compares it to a worker who can do anything you ask, interprets your instruction however it wants, and gives you no way to tell whether it's on the right path.
In older technology work, you got stuck on an install for an hour or a day and had to stop and think. These tools don't get stuck. When they hit a wall, Gary said, they dig a tunnel underneath it, and you may never know how long the tunnel took. Josh tied it to running an MSP without a plan: he built fast for a day with no stated goal, then had to ask the tool to backfill a roadmap for what it had already done.
- Write the goal and the plan before you open the tool.
- Check the work at set points, because the tool won't stop itself.
- Build the roadmap first instead of reverse-engineering it afterward.
Why Do MSPs Have to Take Risks on New Technology?
Because every MSP grew by saying yes to technology it had never touched. Gary pointed to the first time a client wanted a router model you had never configured, and you said yes and figured it out. If someone had checked the work, they would have found problems, but it worked, and you built on it. His concern is that many MSPs now take very little risk, for example reselling cloud servers at a markup instead of understanding how the infrastructure works.
Building software flips that, because you take full ownership of something you may not fully understand. He compares this moment to the late 1990s, when the computer shop he worked at pitched a website to every customer who came in to buy a computer, at $500 to $1,000 apiece, and he admits most of that work was best effort. The web moved fast after that, and anyone who didn't level up was outpaced. The real question is how much risk and liability you are willing to carry for the money.
- Decide how much risk and liability a build is worth before you price it.
- Expect your first version to be outgrown, and plan to level up past it.
- Build skill on internal projects before you take on client liability.
Will MSPs Become Software Development Companies?
Gary thinks many will. His prediction is that over the next 10 years or so, the MSP turns into a software development company that also delivers managed services. Most businesses run broad systems built for millions of users that fit them only roughly, so they bend their workflows to the software. AI makes it possible to build what a client actually needs, sometimes replacing a stack of separate tools with one that does the job better for a fraction of the cost, and the developers building that future use the same tools an MSP already has.
Gary also sees a place for a software development partner who can take over a proof of concept the MSP can't carry further. The fork, as he frames it, is whether you become an AI software development company or keep reselling someone else's AI in the cloud. Josh's practical take: inventory your team and what you've built for yourselves, get uncomfortable internally first, then find one client who understands where you are and be willing to give the fee back.
- Inventory your team's skills and your own internal builds first.
- Line up a development partner for builds that outgrow your team.
- Start with one client who knows exactly where you are with AI.
Frequently Asked Questions
Should an MSP sell software it built with AI?
It can, if it sells the build as a proof of concept to a client who understands what that means. On this episode, Gary Boyle said the hard part isn't the first version but the first change request. Agree before you start on who supports the build, who owns later changes, and what the client can expect from them.
What is the AI danger zone?
It's Gary Boyle's name for the stage where a business, or the MSP serving it, builds automations, agents, and dashboards that work well enough to rely on but aren't documented or supportable. It's dangerous because the work is exciting and easy, and a build can work for a week before failing on the day its builder is out.
What are the levels of AI adoption for a business?
In Gary Boyle's framework, patient zero has heard of AI but doesn't use it. Level one uses it for search, writing, and documents. Level two builds interactive tools, scheduled tasks, agents, and dashboards. Level three needs a real application that other people use, which brings hosting, security, and change management with it.
What does developing in green pastures mean?
It describes building software that nobody depends on yet. With no users and no data that matters, you can change a screen and nothing breaks. Gary Boyle picked up the phrase from a software developer. Once the software belongs to a client, every change has to work without breaking something the client relies on.
Who supports an AI automation when the person who built it is out?
Someone has to be able to, which takes documentation, a repeatable build, and more than one person who understands how it works. Gary Boyle's example is a technician who doesn't show up, leaving a colleague to support a workflow with no documentation while the client's CEO waits on it. At that point the MSP has charged for something that no longer works.
Will MSPs become software development companies?
Gary Boyle expects that over roughly the next 10 years, many MSPs will become software development companies that also deliver managed services, because AI lets them build what clients actually need instead of reselling broad tools. Whether a particular MSP should is a decision about risk, staffing, and the kind of company its owner wants to run.
Episode Highlights
- Building a working tool with AI in a day, without writing code.
- Selling websites in the late 1990s and what it teaches about AI.
- Gary's levels of AI adoption, from patient zero to level three.
- The danger zone: builds that fail when their builder is out.
- Why AI tools never get stuck, and why that's a risk.
- Building first and backfilling the roadmap, in AI and in business.
- Green pastures and the first change request.
- Saying yes to technology you've never touched.
- The fork between AI software development and reselling.
- How to start: internal builds first, then one willing client.
About the Co-Host: Gary Boyle
Gary Boyle is COO and Management Consultant at Bering McKinley. He works with MSP leadership teams on operations, planning, and execution.
About the Host: Josh Peterson
Josh Peterson is the CEO of Bering McKinley and host of The BMK Vision Podcast. Since 2004, Josh has worked with hundreds of MSP owners to build operationally sound, profitable businesses through consulting, peer teams, and direct coaching.
Connect with Josh Peterson on LinkedIn →
Related Resources from Bering McKinley
- How artificial intelligence is transforming MSP software for IT support
- Future-proofing your MSP for the next wave of technological change
- Organizational strategic planning for business performance
Want to Continue the Conversation?
Deciding what kind of company your MSP becomes with AI is a planning question before it's a technology question. Vision gives your MSP a strategic plan, a consultant who holds your team accountable every month, and your numbers tracked against BMK's standards.
%20logo.png?width=1083&height=200&name=BMK%20(just%20text)%20logo.png)