Sell AI outcomes, bill for training, set guardrails.
At an XChange NexGen 2026 panel covered by CRN, MSP leaders said AI revenue is moving beyond resale into consulting, business transformation and billable training. This guide covers what they said and four moves to consider for your practice.

Published November 2, 2026 · ~6 min read
What the panel said
At CRN parent The Channel Company's XChange NexGen 2026 conference in Orlando, Fla., a panel of MSP leaders and security vendors discussed how AI revenue is changing for solution providers. CRN covered the session in MSP Execs See AI Revenue Opportunities Grow Beyond Resale Into Consulting, Services.
Danny Perry, managing partner of Houston-based ITCubed, said the AI opportunity is shifting away from resale and toward consulting and business transformation. Brian Ruschman, president of Covington, Ky.-based C-Forward, said MSPs should be paid for AI training. Jason Myers of OpenText Cybersecurity moderated the panel and raised shadow AI. Nathan Ostrowski of Petra Security urged caution about how much data AI chatbots demand.
According to CRN, ITCubed's project work has increased as clients move past initial assessments and executive projects. That work can create monthly recurring revenue and ongoing agent maintenance.
Why resale is not the business
Perry put the question this way: “Not necessarily being just an IT provider, but how are you transforming how this business works, how it performs, how it increases its revenues, how it reduces its expenses, how it mitigates risk utilizing AI?”
He said MSPs already bring some of this: “And that’s where we all have some degree of expertise in. And we’re gaining more expertise every day.”
CRN reports two practical recommendations from Perry. First, take an outcome-based approach: identify a client's specific problems and see whether AI can help, rather than treating AI as another vendor product to resell. Second, adopt AI internally, so your team can discuss the technology with clients from its own use.
Our reading: the first point changes what you sell, and the second makes the advice credible.
Training is billable
Ruschman said solution providers can't leave money on the table while clients have a strong need for AI training. MSPs, he said, need to treat education as part of the project.
“We always want to say that stuff maybe is part of a sales process, and it certainly is,” Ruschman said. “But training, for sure, we need to be collecting on stuff like that.” He put it plainly: “We need to get paid for it.”
Per CRN, when C-Forward introduces new cybersecurity tools for a client, it conducts billable training every quarter for certain clients. Ruschman said the same goes for AI: recorded sessions, live sessions and lunch-and-learn events can all be revenue opportunities.
MSPs as guardrails
According to CRN, Perry said solution providers will create the governance for the AI operating layer. That means helping clients connect core systems to AI with:
Protocols
Permissions
Auditing
Operational memory
Humans in the loop
“We are the guardrail in so many ways,” Perry said.
Ostrowski raised a related concern. He said solution providers have to be cautious about how much data AI chatbots demand and how far they reach into the systems they touch. He has seen popular chatbots import users' browser cookies. And because AI chatbots use device code flows, stopping device code phishing attempts on items like legacy printers has become more complicated, per CRN.
“Over an exec’s dead body are you stopping those [chatbots] from getting access to an organization,” Ostrowski said.
Shadow AI is already in the room
Myers said shadow AI is a growing concern as users turn to their preferred chatbots. “The users are finding it first, experimenting with it first, solving problems with it first,” he said, and then the business is left to work out what to do about it. His warning to MSPs: “Your customers aren’t starting at zero.” In his words, “They’re already testing tools and uploading documents.”
He noted that organizations are using AI tools to “vibe code” their own software, which can lack the security and governance workplaces need. AI guidance, AI support and AI consulting, even as a service, are options for MSPs, he said. “We all know the cleanup work still shows up,” he said. “But increasingly, the business opportunity shows up, too.”
Ruschman argued for “a happy balance” between shutting every AI tool down and allowing every tool. “You can’t sit back and not use it at all,” he said. He said shutting down all AI will likely lead employees to prompt chatbots on their personal devices. “That’s even remarkably worse,” he said.
What changes for tier-one technicians
Perry said solution providers will continue to need tier-one technicians, but the work will look very different. Technicians will still need certifications. Per CRN, they will also need to understand new AI systems, how the MSP works holistically, the outcomes agreed with customers, and how to prompt and guide AI systems to perform work.
Ruschman said tier-one technicians already look remarkably different than they did in 2006. He expects them to use more business data and do more double-checking of work automated through AI agents. He sees legal decisions and permanently deleting data staying away from agents. “There’s going to be a while here where any AI agent, any tool deserves a check in a lot of ways,” he said.
Ostrowski said AI models are “spikey” and show extraordinary capabilities in specific tasks while failing at others, so MSPs can help users understand when to apply certain models. “They think it’s a uniform blob. But you know where the spikes are.”
Four moves for MSP owners
These moves are Consultiply's reading of the panel. They are not recommendations from CRN or the panelists.
Start from a client problem
Before naming a tool, write down the specific problem, the outcome the client wants and how both sides will know it worked. Use AI in your own operations first.
Scope training as project work
Put AI training in the project scope as its own line, whether it is recorded sessions, live sessions or lunch-and-learn events, instead of giving it away during the sale.
Write the guardrails down
For each client, document how AI connects to core systems, what it is permitted to reach, what is audited, what the system keeps as operational memory and where a person approves. Keep legal decisions and permanent data deletion with people.
Start with the AI already in use
Before drafting a strategy, ask which tools staff already use and which documents they upload. Aim for the balance Ruschman described, between banning everything and allowing everything.
Read with care
These are opinions from one conference panel, as reported by CRN.
They are not survey or market data.
Two speakers work for vendors.
Myers is cybersecurity strategist for MSP partnerships at OpenText Cybersecurity, and Ostrowski is chief operating officer and co-founder of Petra Security. Their views may reflect their companies' interests.
CRN's parent, The Channel Company, runs XChange NexGen, and both MSP panelists are CRN MSP 500 honorees.
Ruschman cited data on employee AI tool use without naming its publisher.
This guide does not repeat it.
This guide is Consultiply's independent reading of CRN's panel coverage. It is not affiliated with or endorsed by CRN or the panelists.
Reviewed, not run by us.
Consultiply's independent reading of CRN's panel coverage; not affiliated with or endorsed by CRN or the panelists. CRN, XChange NexGen 2026 coverage (October 2026) →