ConnectWise Urges MSPs to Trade AI Promises for Measurable Results
- Anjali Fluker
- 883 views
ConnectWise executives urge MSPs to standardize their tech stack, pilot one repetitive task in shadow mode and measure results before scaling AI.

Every MSP has heard the pitch: AI will cut costs, speed up operations and let you take on more clients. The hard part is working out which tools deserve a slot in your stack. ConnectWise's answer is to build AI into the platform MSPs already run, so the work of adoption happens inside tools that are already in place. At IT Nation Connect Global 2026, CEO Manny Rivelo and Chief Product and Technology Officer David Raissipour laid out four practical steps for turning that promise into operational results.
The core idea is simple. AI should take the routine work off technicians' plates so providers can spend their hours on what customers actually buy. Rivelo frames it as a core versus context question. "What is core to an MSP is what differentiates them," he told ChannelPro. "If I'm an end customer, I want to buy the service from you. It's not because you sell tickets."
Standardize the stack before you automate
Automation depends on data that can move between systems. Rivelo recommends that MSPs start by reviewing the tools they use to run the business, then consolidating where they can. A leaner stack makes it easier to connect information, automate workflows and apply AI across operations.

Manny Rivelo
"I'd standardize as best as possible to get common data, a reasoning engine and a system of record," Rivelo said. "You could do all that by thinking about your tech stack."
ConnectWise is building agentic capabilities directly into its platform rather than selling automation as a separate product. The company says that approach reduces the integration work MSPs need to do before they can adopt AI. Raissipour also described build-your-own-agent capabilities that let partners create agents for their own business needs. Those agents run on the ConnectWise agentic framework, which includes integrations, monitoring and governance controls.
Start with one repetitive task
You don't need to automate the whole business to see value. Rivelo's advice is to pick one routine activity that eats a lot of technician time. Automating credential resets is one example he points to, since it can cut ticket volume and free technicians for complex customer issues. Early wins like that also help build internal confidence. "Then, use that as a vehicle to educate your MSP and change the culture," Rivelo said.
ConnectWise's Agent Lab is meant to lower the barrier to that first step. Customers on the platform can choose prebuilt agents trained for their environments. The more interesting feature for cautious teams is shadow mode. An agent in shadow mode reports what it would do without actually doing it.
"The implementation timeline is really simple," Rivelo noted. "If you could put the agent in shadow mode, it will tell you what it's going to do but won't actually do it."
Technicians review the proposed actions and decide which ones to approve. Once the agent's track record looks solid, the team can turn shadow mode off and let it run with more independence. For MSPs that are wary of handing over control, that staged rollout is a sensible way to build experience first.
Automate responsibly, not just quickly

David Raissipour
As AI tools get easier to build, there's a temptation to create a custom agent for every problem. Raissipour warned that building an automation is only the first part of the job. Security, governance, ongoing maintenance and scalability matter just as much, and he pointed to AI developers who ship custom agents without the infrastructure to support them over time. That gap can create risk for MSPs and for their customers.
"I sincerely believe that people are underestimating the risk associated with AI that is not governed, monitored or carefully planned," Raissipour said.
His advice is to start with the business problem instead of the tool. "Design for the outcome, not for the tool that you have in front of you," he said. In practice, that means defining what needs to improve, deciding how you'll measure success and then picking technology that can deliver it. It also means accepting that building an AI solution in-house isn't always the best use of an MSP's time. ConnectWise's answer is a framework partners can use for their own automations while leaning on the platform's underlying capabilities.
Turn efficiency into better service and pricing
A common worry is that AI will push prices down or eliminate jobs. Raissipour argues the better question is what MSPs can do with the time automation frees up. Technicians can move toward work that requires judgment, creativity and a real understanding of customer needs. "The goal is to figure out how many people can we free up to go solve the next problem that I can't automate and deliver more value," he said.
The service gains are concrete. MSPs can improve response times, strengthen service-level agreements and, in some cases, prevent problems before customers notice them.
- Improve response times
- Strengthen service-level agreements
- Potentially prevent problems before customers experience them
Over time, those gains could change how MSPs charge. Rivelo said the traditional endpoint-based model may give way to outcome-based pricing as automation matures. "What if I said, 'You, Mr. Customer, have 50 or 100 tickets a month. If I can get that down to five tickets a month, would you pay me more?'"

Ann Westerheim
Preventing repeat issues and cutting customer downtime is a better measure of value than ticket counts. It also gives MSPs a way to compete on service quality rather than on price alone. Ann Westerheim, founder and president of Greater Boston-area MSP Ekaru, sees the same opportunity. She told ChannelPro that ConnectWise is going all in on AI, with capabilities including PSA assistants, intelligent ticket intake and agents for dispatch, backed by aggressive trials and pricing. She also noted a consistent message across the conference that MSPs investing in these tools are seeing stronger operational performance and profitability. "The big question now is how MSPs turn that promise into measurable results," she said. "The opportunity is clear, but successful execution will be what separates the leaders."
Measure first, then scale

Nobody expects MSPs to transform overnight. Still, both executives argued that AI is moving fast enough that waiting carries its own cost. The practical sequence is short: find a recurring problem, test whether automation can handle it and measure the result. Once that loop works, expanding to harder tasks becomes a matter of building on proven results rather than making a leap of faith.
My read is that the real test for MSPs won't be how many tickets an AI agent can close. It will be whether technicians spend less time on routine work and more time on the expertise customers will pay for. That's the measure worth tracking from the first pilot onward.