The AI Line in the Plan Is Mostly a Pilot
By Gio OlavarriaFourteen percent of MSPs run an AI service line that brings in revenue. That is the count from a survey Nerdio published on October 20, 2026, of 153 MSP decision-makers. A month earlier, Informa’s 2026 MSP 501 survey insights found that 57 percent of MSPs expect AI to rank among their biggest revenue gainers over the next twelve months, up from 37 percent the year before.
The distance between those two numbers is where the AI line in most value-creation plans sits. The line shows a forecast. Behind it there are usually a few assessments and some resold licenses. My argument: until the AI line can be shown as a row on the P&L, with clients under contract, a price, a margin and renewals, it is a pilot, and the plan should carry it as one.
I run revenue. I am also skeptical of most AI claims in this industry, which is a bias the reader should know about. The numbers below are the industry’s own.
Expect, offer, earn
Kaseya’s 2026 State of the MSP report surveyed 1,061 MSPs in November 2025, and it asked about AI three different ways. The answers fall at each step.
Almost half, 48 percent, named AI and automation the top client need for 2026, ahead of security. Thirty-eight percent said they offer AI and automation services. Thirteen percent named AI and automation among their top three revenue sources. Kaseya’s own explanation for the gap is that “many providers are still defining, packaging and pricing these services.”
Read those as three different claims. The first is what MSPs believe clients want. The second is what MSPs say they sell. The third is where the money comes from. Most AI talk in a management presentation is the first claim dressed as the third.
Nerdio’s survey, fielded this summer, lands in the same place. Thirty-two percent of the MSPs it asked have adopted AI in their own operations, 14 percent run a revenue-generating AI service line, and 12 percent can point to measurable productivity gains. Two cautions belong with these numbers. The sample is small and limited to MSPs that manage Microsoft 365 and Azure for clients. And Nerdio sells a platform for exactly that work, so its conclusion, that the MSPs with standardized operations will win, is also its sales pitch. Neither survey publishes a price, a margin or a count of contracts behind the AI lines it found.
What an AI line usually holds
Picture an AI line in a plan. It might include readiness assessments sold to a dozen clients, Copilot licenses resold through a distributor, a governance policy offered to compliance clients, and a curve that climbs from there.
Each piece is a different kind of revenue. An assessment is a one-time project. A resold license is Microsoft’s product at the distributor’s margin. It would sell whether or not the MSP had an AI practice, and it carries none of the delivery that makes managed services worth a multiple. A governance policy can be real recurring revenue when it rides on an existing compliance agreement. The curve is the 57 percent: an expectation.
The line still has value. It is a set of small bets that have not yet proven they renew. A plan that books it at the same confidence as managed security revenue is mixing a forecast with a contract.
What buyers pay for
The buyers have been explicit. Drake Star’s Q2 2026 MSP update, published in July, says buyers “pay for MSPs that have embedded AI into service delivery, not those merely selling it,” and that “buyers are no longer paying for AI claims, they’re paying for AI proof.” Craig Fulton, an M&A adviser at Evergreen, told Channel Insider in August that AI has not yet produced a separate valuation premium. In his view, AI that improves scalability, efficiency and profitability shows up in the financial measures buyers already check.
That moves the AI question from the revenue line to the margin of the lines the MSP already sells. If AI cuts the technician hours behind a help desk contract, the profit on that contract goes up and the buyer pays for it at the multiple that contract already earns. Kaseya’s report says MSPs “remain in the efficiency phase of AI adoption,” applying it mostly to monitoring, alerts, security and tickets. That is the right place to look for value. It is also where Nerdio found that only 12 percent can show a measurable gain.
So the owner has two separate claims to test: the AI revenue line, and the AI effect on delivery cost. Each needs its own evidence.
Five questions for the AI line
These go to the CFO and the head of services before the AI line goes into the plan, and again at each operating review.
- Which clients pay for an AI service under a signed agreement, and how many have renewed it at least once?
- What is the price basis (per user, per device, fixed fee, usage), and is it written in the service catalog, or quoted deal by deal?
- What is the gross margin on the line after every vendor cost, including licenses, model usage and the hours of the people who deliver it? Microsoft’s move to usage billing for new Copilot Business licenses, starting December 1, makes the last part of that question harder to answer from a flat per-seat price.
- How much of the line is a vendor’s product resold, and how much is a service the MSP delivers itself?
- For the services the MSP already sells, what has AI measurably changed: tickets per technician, time to resolve, hours per client? Measured against what baseline, and over how many months?
A line that cannot answer the first question is a pilot. Carry it in the plan as an option with a cost, give it a date to prove itself, and leave it out of the revenue bridge that sets the exit price. A line that can answer all five is a service line, and it should be priced and reported like the others.
The row
I could not find a single MSP in 2026 that published its AI service revenue, its client count on AI contracts or a margin on the line. I could not find a diligence provider that publishes how it classifies AI revenue in an MSP deal. The industry is forecasting a revenue line it has not yet shown.
That leaves the owner with the cheapest test in the value-creation plan: ask for the AI line as a row on the P&L, with the clients, the price, the margin and the renewals beside it. If the row is there, price it like the rest of the business. If it is not, the AI line is a pilot, and the plan should say so.
Frequently Asked Questions
How many MSPs actually earn revenue from AI services?
About one in seven, by the two most recent counts. Nerdio's survey of 153 MSP decision-makers, fielded June 22 to July 7, 2026 and published October 20, found 14 percent operate a revenue-generating AI service line. Kaseya's 2026 State of the MSP report, 1,061 MSPs surveyed in November 2025, found 13 percent identify AI and automation as a meaningful revenue source when asked to pick their top three. Neither survey publishes a price, a margin or a contract count for those lines, and Nerdio sells a Microsoft 365 and Azure management platform.
Do buyers pay a higher multiple for an MSP's AI revenue?
No, on the evidence published so far. An Evergreen M&A adviser told Channel Insider in August 2026 that AI has not yet produced a separate valuation premium, and that AI used to improve scalability, efficiency and profitability strengthens the financial metrics buyers already assess. Drake Star's Q2 2026 MSP update says buyers pay for MSPs that have embedded AI into service delivery, not those merely selling it, and that buyers are now paying for AI proof.
What should an owner ask before an AI line goes into a value-creation plan?
Five things. Which clients pay for it under a signed agreement, and how many have renewed. What the price basis is, and whether it is written in the service catalog. What the gross margin is after every vendor cost, including licenses and metered usage. How much of the line is a vendor's product resold and how much is a service the MSP delivers. And what AI has measurably changed in the cost of delivering the services the MSP already sells. A line that cannot answer the first question is a pilot, and it belongs in the plan as an option, not as revenue.