← Back to Blog CHANNEL

Patrick Leonard on building Backup Radar — and what he changed for MSP Process

An interview with our founder on the lessons from running a 15-year MSP, building one category-defining product, and starting over for the channel's next problem.

Patrick Leonard founded Backup Radar in 2014 after operating his own MSP for 12+ years. Thousands of MSPs use Backup Radar today. After a successful exit, he started over — to solve the helpdesk identity problem the industry was avoiding. We sat down with him to talk about what he learned the first time, what he changed for MSP Process, and what he thinks the channel still gets wrong.

What was the moment you knew you had to build Backup Radar?

I was running my own MSP. Backups were the source of every late-night call, every awkward client conversation, every CYA email I had to send. I kept thinking: this is a reporting problem, not a backup problem. The backup software was fine. We just had no consistent way to know whether it had run, whether it had succeeded, whether the data was actually recoverable. So I built the reporting layer for myself. Then a peer asked to use it. Then ten peers asked. That's how it started.

What's the biggest thing you got right with Backup Radar that you carried into MSP Process?

Two things. First — the channel doesn't want a tool that competes with their other tools. They want a tool that makes their existing stack better. Backup Radar didn't try to replace ConnectWise or Veeam. It made them work better together. We built MSP Process the same way. We integrate with ConnectWise, Autotask, HaloPSA, ServiceNow, Zendesk, Freshservice — we don't replace them. We make them more verified, more orchestrated.

Second — solve a problem the channel is already worrying about, not one you're trying to convince them they should worry about. Backups were already a sleepless-night problem. The helpdesk vishing threat is the new sleepless-night problem.

"I'm not interested in convincing MSPs they have a problem. I'm interested in giving them the tool when they already know they do."

What was the moment you knew you were going to start MSP Process?

I was watching the MGM and Caesars breaches roll out, and the post-mortems were almost identical. A phone call, an unverified password reset, ransomware deployed within hours. I started asking peer MSP owners how they were handling helpdesk identity verification. The answers ranged from "ask the manager's name" to "we kind of don't." There was no platform solving this. The single biggest attack vector against MSPs and the clients we serve, and there was no answer.

Then I started talking to my insurer, and they were asking the same questions. The cyber-insurance market knew the problem before the security industry did. That tells you everything.

What's different about building MSP Process versus Backup Radar?

Backup Radar started as a reporting layer on top of someone else's data. MSP Process is a platform that has to do real work in real time — answer voice calls, verify identities, orchestrate workflows across seven channels, integrate bidirectionally with PSAs and ITSMs. The technical surface is enormous compared to Backup Radar. We have AI Voice, multi-channel ingestion, identity verification, automation bots, scheduling, on-call. Each one of those would be a startup on its own.

The other difference: I built Backup Radar mostly alone in the early years. I'm not building MSP Process alone. Greg Celmainis runs operations and corporate development. Chris Reid leads product and engineering. The team has been MSP operators, dispatchers, technicians. Every product decision goes through someone who's lived the problem.

What's the most common pushback you get from MSPs?

"We can't replace our helpdesk with AI. Our clients want to talk to humans."

I love this objection because it's based on a misread of what we do. We don't replace humans. We replace the unverified intake. AI Voice answers, verifies, triages, and routes. If the caller wants a human, they get one — but a verified, contextualized one, with the ticket already open and the issue already classified. The tech who picks up doesn't have to spend the first 5 minutes asking "who am I speaking with."

The other version of the pushback: "We don't get vishing attacks." That's almost always wrong. They just haven't detected one. We've had customers catch attempts in their first 30 days that they had no idea were happening.

What do you think the channel still gets wrong?

The channel undersells itself on regulated clients. Healthcare, finance, legal, defense — those are the clients with the highest willingness to pay, the longest contract terms, the strongest renewals. And most MSPs avoid them because they don't have the SOC 2 evidence pack ready. The platform should give you that. It should be the natural exhaust of running the desk.

The other thing the channel undersells: AI Voice as a revenue line. Every MSP can resell white-labeled AI Voice to their own clients. 40–67% gross margin per seat. We've seen MSPs replace half their backups-and-monitoring revenue with AI Voice over 18 months.

The take-home

Build for the channel. Integrate, don't replace. Solve the problem the operator already has, not one you're trying to manufacture. And take the regulated business — that's where the renewals live.

Patrick still spends a day a week in the support queue at MSP Process. He says it's the only way to know what's actually going wrong. We believe him.

Ship a verified service desk in 30 days.

Book a 30-minute call with a solutions engineer who came out of an MSP service desk. Bring your stack. We'll model the impact with your numbers.