This is the first in a new series looking at how Manage Protect’s partner MSPs are putting AI to work: building integrations, automating processes, surfacing data and helping their teams service clients more efficiently.
Linden Jackson, Founder and MD of Agile IT, didn’t set out to become an AI power user. He started, by his own admission, because he’s “lazy and like[s] to have other people do stuff” for him. Several months on, Claude is running dashboards, has rebuilt the entire website and is quietly reshaping how the Agile team handles support calls.
A Saturday morning dashboard
It started as a bit of a brag session in Linden’s peer group. Other MSP owners were talking up their integrations, and Linden decided to see what he could do himself. He opened Claude and asked for one thing: a dashboard pulling data from HubSpot, Halo PSA and Zomentum.
Working read-only to keep things safe, he cross-referenced client, ticket and device data across systems, something none of the individual platforms could do alone. By Monday’s peer group meeting, he had a working dashboard to show for a weekend’s effort.
“I think that AI has really opened up opportunity for people to get things started and then to keep the momentum going. I'm one of those people who probably get stuck along the way and then I'll get distracted and never get it finished. And now suddenly I'm finding ways to get things moving.”
A website built by conversation, not WordPress
The bigger transformation came when Linden fed Claude his brand documents, old website and target client profile, then simply started talking it through changes.
The result: a hand-coded HTML site (no CMS, no plugins, no monthly fees beyond hosting) with 70-plus pages, an interactive service diagram, a pricing calculator, a hardware-replacement calculator and a self-service security health check that scores visitors and feeds results straight into HubSpot.
“It’s like I’m in a room with an expert web developer… I can just go, hey, can you move that logo there? Can you put that picture there?… Now, if they were a human developer, they would have got totally annoyed with me and gone home, but being an AI: oh yeah, cool, we can do this.”
Beyond the website: dashboards and smarter support
One of the most practical builds has been a middleware layer connecting Agile IT’s phone system to Halo. When a call comes in, it matches the caller to their Halo record, surfaces open tickets, transcribes and summarises the call into a private note, tracks sentiment, and lets staff add a new contact in one click if there’s no match.
“It’s been able to go, right, this is the person we’ve matched to the database. Here’s their open tickets. Is this call about one of these tickets or is it about something new? If it’s about something new, create a new ticket. It then takes an AI summary of that call and puts that into a private note in the ticket.”
Lessons for other MSPs
Linden is candid that not everything he’s built has been worth it. His early dashboard experiments duplicated functionality other tools already had, and he’s wary of the gap between what a technically capable business owner can build and what the rest of the team can actually use and troubleshoot.
His advice for MSPs starting out:
- Go AI-first. “Everything that we are doing needs to start to become AI-first thinking” — evaluate every tool purchase on whether AI meaningfully improves it.
- Judge tools on data access, not features. “Can I access the data that I need from this system? …how do those tools talk to each other?” matters more than any built-in report writer.
- Stay inside your competence. Read-only first, tight API scopes, and know where to stop, the same way you wouldn’t start servicing your own car with a YouTube video.
- Don’t let AI just agree with you. It’s designed to be agreeable, so tell it explicitly to validate and challenge your ideas rather than rubber-stamp them.
- Build for the team, not just yourself. A tool only the founder can fix isn’t actually solving a problem for the business.
"Everything that we are doing needs to start to become AI-first thinking"
The next challenge
Agile IT offers an AI support package, but take-up so far has been slow. Linden sees a similar pattern playing out to cybersecurity, where some clients now simply expect their MSP to be across AI without necessarily paying for a separate line item. Packaging and selling that support is still a challenge and may just be a matter of client maturity and education.
Linden also sees a bigger risk: AI is becoming another front on which MSPs need to defend the client relationship. AI vendors and platforms are increasingly going straight to end users with their own shiny tools. If MSPs aren’t the ones having the AI conversation first, clients may simply try to figure it out on their own, without the guardrails or judgement that Linden has been careful to apply throughout his own experiments.
While dashboards, websites and support tools are AI helping humans work faster, agentic AI is the next level. This is a shift Linden sees as a whole different ballgame, and one that will profoundly impact how we work.