Agentic AI is entering a new phase for managed service providers, with early deployments shifting from chatbots toward practical network automation, according to Auvik President Mark Ralls.
Rather than replacing engineers, Ralls says the technology is helping MSPs reduce operational overhead, improve technician efficiency, and proactively resolve network issues.
How AI adoption is maturing across MSPs
Ralls said adoption among MSPs still follows a traditional technology adoption curve, with only a small percentage aggressively deploying AI-powered workflows today.
“We’re seeing your typical tech adoption curve,” Ralls said. “A small handful are really going hard and doing some really impressive things… but I think that’s like five or ten percent, perhaps at the outside.”
Most solution providers, he said, remain in an evaluation phase, increasingly influenced by customer demand for AI-enabled operations.
“I was talking to the president of an MSP earlier today who said they’re following what their customers are asking for,” Ralls said. “His customers seem more willing to have automated resolution from various tools than perhaps he thought they might be.”
Why Auvik is focusing on operational AI instead of chatbots
The discussion comes as AI continues to dominate vendor roadmaps across the channel.
While many technology providers initially focused on adding chatbot-style interfaces to existing products, Auvik is betting that autonomous agents delivering measurable operational outcomes will prove far more valuable for MSPs.
“I think there was a lot of early hype,” Ralls said. “Two or three years ago, it was, ‘I’m just going to slap a chat interface on what was already a perfectly functional product.’ It’s not clear that actually made anything better.”
The shift to agentic AI built into the platform
Instead, Auvik has focused on AI agents embedded within its network management platform to streamline existing workflows.
Examples include AI-assisted alert creation, root cause analysis, intelligent alert filtering, and Network Terminal Assist, which helps technicians generate vendor-specific command line instructions using retrieval-augmented generation (RAG) built on Auvik’s proprietary network expertise.
“The whole idea is giving our partners superpowers,” Ralls said. “Right now, our focus is on how do we make it easier and more streamlined for partners to do the things they’re already doing. It’s much more about making the status quo less painful rather than taking a future-state approach.”
Agentic AI helps address staffing and alert fatigue
Rather than replacing engineers, Ralls argues that agentic AI can help MSPs maximize limited technical staff while improving employee satisfaction.
He pointed to one early beta customer whose Tier 1 technician resolved an issue independently instead of escalating it to senior engineers.
“Our Tier 1, who was on call, did not have to page Tier 2 or Tier 3,” Ralls recalled from customer feedback. “They were able to take the steps the AI recommended and resolve it.”
That approach addresses one of the industry’s ongoing challenges: hiring and retaining experienced networking professionals while managing increasingly complex customer environments.
“When your Tier 1 can resolve something automatically, your Tier 2 or Tier 3 doesn’t get paged on a Saturday afternoon. It’s hard to put that on an income statement, but it leads to a better quality of life and lower attrition,” said Ralls.
Alert fatigue reductions provide focus on important notifications
The same philosophy applies to alert management, where AI can reduce operational noise by filtering non-critical notifications and allowing technicians to focus on issues requiring human expertise.
“AI allows us to work toward a point where anything that hits a human’s desk is something that actually needs a human, and the rest of the paperwork and distraction is handled behind the scenes,” Ralls said.
He added that no technology is perfect, and a small number of false positives do make it to a technician’s desk.
Still, as alert fatigue continues to grow, any reduction in the noise across an environment is a substantial win for most MSP teams.
Autonomous network remediation is the long-term goal
Ultimately, Auvik’s roadmap extends beyond assisting technicians toward proactive network automation.
Ralls described a future where AI agents detect developing issues—such as abnormal CPU utilization—predict likely outcomes based on historical data, and automatically apply proven remediation steps before users ever notice a problem.
“The network should just work,” he said. “When you bring in a partner that uses Auvik, network issues are going to be resolved before anyone notices.”
“The whole idea of a managed service provider is ensuring customers can operate their business without worrying about the headache that is IT. How do you do that most effectively? You proactively identify potential problems and resolve them before an end user ever notices,” Ralls added.
While fully autonomous operations remain limited today, Ralls expects meaningful progress over the remainder of 2026.
“My crystal ball says at least 25 percent of MSPs by the end of the year have at least a handful of use cases that they trust an AI agent to resolve without a human at all,” he said. “It’ll be basic things, but once we get that comfort level, those use cases will grow steadily in both number and complexity.”