How IT-noch Is Building an AI-First MSP, Starting With OpenFrame

Raymond Alexander leads IT-noch, an MSP with about 700 endpoints across schools, nonprofits, law firms, accounting firms, transportation companies and manufacturers. He's taking the business in a new direction: "We're actually trying to transition from being your traditional MSP to more of an AI architect type of firm."

He spotted OpenFrame on Instagram, joined the beta and liked what he saw. Getting it running took "not long at all." He pushed the install script out through his existing RMM, and scripts on OpenFrame run immediately.

At the interview in February 2026, it was early days. OpenFrame ran alongside his existing RMM, PSA and EDR tools. His team hadn't yet changed the way it works to use Fae and Mingo, so its workload hadn't changed. He still sees a lot in it: "That AI feature that you guys have could be game changing."

IT-noch doesn't stick to one niche. "We work with all verticals," Raymond says. His team looks after about 700 endpoints for schools, nonprofits, law firms, accounting firms, transportation companies and manufacturers.

The business is changing shape. Raymond wants IT-noch to spend more of its time inside client organizations, working out where AI fits, with managed services as the second line.

His support model follows the same idea. In his words, "The AI should take care of level one, maybe level two." His technicians take level three and the work that needs someone on site. He's already working toward that model.

End users are the other piece. Today they get support in one of two ways: they log a ticket or they pick up the phone. An AI agent on their own computer is new to them. As Raymond puts it, "So it's going to be a learned behavior for them to utilize Fae."

Raymond found OpenFrame on Instagram, signed up and got into the beta. It was still early, but he could see where it was heading: "I looked at the potential of it and was like, man, this actually could be pretty cool."

Setup was quick. Asked how long it took to get running on real endpoints, he said, "Not long at all." He put the OpenFrame install script into his existing RMM and pushed it out. Scripts on OpenFrame run immediately, which he likes.

Fae, the AI agent that sits on the end user's computer, is the part that caught his eye: "Having an on-site tech right there for someone is actually pretty intuitive." His end users ask Fae general questions. At the interview, Fae couldn't yet resolve a lot of their issues, and when it couldn't, Raymond picked up the ticket himself.

Mingo, OpenFrame's technician copilot, wasn't part of the routine yet. Raymond wasn't sure how work moves from Fae to Mingo, or how to hand Mingo a ticket. He knows the order he wants: Fae answers what it can and tells the end user when it's raising a ticket, Mingo picks that ticket up first, and a person steps in after that.

He also likes the open-source model: "I could utilize any AV or any third-party tool because it is open source." He hasn't tied one in yet, but he knows where it leads: "I know once I learn it, I see the potential."

At the interview in February 2026, OpenFrame was running on his endpoints, deployed by script through his existing RMM in "not long at all." Scripts ran immediately. His end users were starting to ask Fae general questions.

Raymond was clear that it was too early for results. Asked whether his technicians' workload or efficiency had changed, he said no. His team hadn't changed the way it works to use Fae and Mingo: "I see the potential. It's just that we're not utilizing the tool."

His view of the platform comes with conditions attached. "That AI feature that you guys have could be game changing," he says. And at the close: "I really do like the platform, and I think that you guys are on to something."

What he still wants, framed as polish: email-to-ticket, so end users can raise tickets by email and OpenFrame can work as a full PSA; a clear path from Fae to Mingo to a person, with Fae questions kept apart from tickets and a clear answer when a chat says it's "waiting for approval"; an introduction to Fae for end users when it's deployed; and guidance on what Mingo can do and how to hand it work. In his words: "Polishing Fae and Mingo and allowing for email, making it a true PSA, you do that, you'll have a corner on the market."

What he expects next: once his team changes the way it works around Fae and Mingo, he'll be able to say what has changed. "I'm really excited about utilizing the final product."