From Experimentation To Execution: How MSPs Are Using AI Agents Today
‘Once you actually get into automation and business process flows, to bots, to apps—then you’re really going to help that company move forward. And as you look at SMB customers, they need that edge because a lot of them run so lean,’ says Abby Hanson, chief digital business officer at Razor Technology.
AI agents are actively reshaping how many MSPs run their business, support customers and think about scale. Some MSPs are moving cautiously with agents, as others are moving more aggressively.
For Shayon Mazumder, AI agents are already embedded in day-to-day operations. Mazumder, managed IT services practice leader at MRE Consulting, said AI-powered chat and voice agents are being used to improve first-call resolution and help service desk teams close more tickets. “It’s transformed our entire business,” he told CRN.
Don Monistere, president and CEO of General Informatics, said interest from customers has surged as AI agents begin to operate with greater autonomy. At the same time, the increased use of agents has prompted the company to rewrite internal policies and create isolated testing environments for employees experimenting with new tools.
Not every MSP feels ready, though. Jason Null, CTO of MIS Solutions, described agentic AI as both inevitable and unsettling. With limited controls and governance in place, he said the sudden availability of AI tools has created a “Wild West” scenario where MSPs are still trying to understand how to support and secure technologies customers are already asking for.
Still, many MSPs see AI agents as a growth opportunity, especially for SMB customers.
“We recently just did this with a prospect,” Abby Hanson, chief digital business officer at Razor Technology, told CRN. “We brought an agentic AI prototype to the table to show them what it could do to move their business forward. Once you actually get into automation and business process flows, to bots, to apps—then you’re really going to help that company move forward.”
At CRN parent The Channel Company’s XChange March conference in Orlando, Fla., this week, CRN spoke to a handful of MSPs about how they’re using AI agents in their own business and the impact they’re seeing.
Check out their answers below.
Shayon Mazumder
Managed IT Services Practice Leader
MRE Consulting
Houston
It’s transformed our entire business. We’re using chatbots, we’ve created our own voice AI agent to help with first-call resolution, and we’re on a journey to understand how we can create more efficiency by using some of these emerging technologies. We use AI assisted triage, chat and voice to help all of our service desk agents resolve as many tickets as possible.
Don Monistere
President, CEO
General Informatics
Baton Rouge, La.
Agentic AI is massively impacting us on all fronts, both internally and externally. We sell an artificial intelligence platform that is kind of an aggregation of three or four toolsets that we use. We’ve done some of our own custom development as well for AI, but we’re getting a lot of questions now about agentic AI and the actual transactional behavior, if you will, of the AI without a human in the loop that sometimes creates some security risks that we get concerned about.
Four weeks ago, it’s like everything changed when OpenClaw made their big announcement and rolled out OpenClaw. Suddenly you have people downloading software that they probably have no business putting in enterprise environments. That’s been our biggest concern and impact initially, and not only for our own internal staff. We want to make sure that we’re not downloading that and putting it in the enterprise network. So we actually rewrote our policy and provide our people with their own environment that was far and away from our normal environment because we wanted to let them test it, obviously, but not within our corporate environment.
The other thing that I’ve seen in the market is more and more people are asking questions about agentic AI. Before it was, ‘We know we need to use AI. I’ll get to it eventually.’ Now people are saying, ‘Oh, wait, I can get a machine to do something that a human can do, and I don’t have to increase my staff? I want to hear about that.’ And so people are asking that question.
And quite honestly, Microsoft’s making a lot of progress in their Copilot product too. That’s starting to get a little bit of traction. You can create agents within their current application. If you’re part of their Frontier program, they already have prebuilt agents that you can utilize for research and various other things. And it all follows the identity profile of the user, so that’s pretty helpful.
We’re getting a ton of questions. In fact, I had a customer call me about Claude Cowork, and she asked me a question I hadn’t been asked yet, which was, ‘Is the business version truly as safe as they claim it to be?’ And I said, ‘Look, I’ve read all the marketing information, but I haven’t tested it, and I can’t make a recommendation until I test it.’ And so I feel a little bit behind, and I’m usually ahead of everything. So when I get back from XChange, I’m going to be testing Claude Cowork over the weekend to see if all of the log file information they provide allows us, as security analysts, the ability to go back and see what the application is actually doing. So no rest for the weary.
Jason Null
CTO
MIS Solutions
Cincinnati
AI in general is a pain point. I don't know how to answer about agentic AI because it’s one of those things that we’ve been hearing about, and at this point in time, we don’t even have any controls around it. I’m starting to look at ways around how can I stop other AIs from being used in customers so that people aren’t just running out to ChatGPT and stuff like that and instead making them use the AI that they’re assigned, like Copilot, and how do I do that with data governance around it. That’s stuff that we haven’t even started to really pursue. We have sold, in the last probably 90 days, three licenses. We sold no other licenses until this point, besides us using it internally with our Copilot licenses. So I don’t have an answer to what exactly it’s going to look like. It kind of scares me a little bit. Our techs use AI, but I don’t know if they understand how to truly support it and how to answer people’s questions. So to answer your question, I don’t know.
We’ve been hearing about it, and all of a sudden now it’s here, and people are finally starting to want to use it. And it’s not like we've had any kind of true training or governance or regulation. I feel like it’s the Wild West still, and we’re not ready for it at all.
Abby Hanson
Chief Digital Business Officer
Razor Technology
Conshohocken, Pa.
As we move into agentic AI, we’ll bring new ideas to the table that they haven’t thought about to be able to help move their business forward. It’s all about how does AI drive ROI within their organization? We recently just did this with a prospect. We brought an agentic AI prototype to the table to show them what it could do to move their business forward.
Once you actually get into automation and business process flows, to bots, to apps—then you’re really going to help that company move forward. And as you look at SMB customers, they need that edge because a lot of them run so lean. We’re driving adoption as well. We’re not just going to say, ‘Hey, here’s your AI.’ We’re actually going to show them how to use it.
Tommy Vaughan
President
Central Technology Solutions
Lynchburg, Va.
We use agents to build workflows. We didn’t focus on how to make our back-end operational system better because I don’t believe that AI is going to replace people. Instead of trying to figure out how to [use] a Large Language Model to make your MSP more efficient with ticketing, imagine if you used AI to be another person within your organization to then manage your processes? You can [take these processes] and tell it, ‘I want you to do that once a month.’ So now I’ve just saved an hour, maybe two hours. What can I do with that? Now I can be more effective on doing those other things within my MSP. That’s where I think the impact’s going to happen. That’s what is going to make us more efficient.
Chelsea Skinner
President, CEO
Oversee My IT
Lewisville, Texas
We have one staff member that’s solely dedicated to just building AI agents right now. His first project is building a marketing agent, and then he’s going to be working on a help desk for the team. Once he has built that, and he understands it, then we’re going to start offering it as a service. We have a lot of SMB clients, and they love AI. And so we would rather get involved early before it derails their business or they use the wrong service and they don’t have the guardrails.