MSP COO: AI Should Turn Knowledge Workers Into ‘Wisdom Workers’

‘The old job was knowing the answer,’ Nathan Phinney, chief operating officer of AllSafe IT, says. ‘The new job is judging the answer.’

While some people fear that artificial intelligence will eliminate jobs, using the technology correctly can turn solution provider employees from knowledge workers into wisdom workers, Nathan Phinney, chief operating officer of AllSafe IT, said.

For solution provider employees, the mindset shift is seeing themselves as decision-makers who can leverage an AI agent’s unlimited attention and knowledge, Phinney told a room of solution providers during CRN parent The Channel Company’s XChange NexGen 2026 event. His Pasadena, Calif.-based solution provider is a 2026 CRN MSP 500 honoree.

“The old job was knowing the answer,” he said. “The new job is judging the answer.”

Nigel Ziyad, CEO of Saint Charles, Ill.-based solution provider Ziyad Technical Resource Consulting, told CRN in an interview that he plans to use Phinney’s AI success stories when discussing with clients where and when to use the technology.

He hopes that helping customers understand AI’s power and risks means “they won’t go off the rails and think, ‘OK, yeah, I’m getting the right answer from AI,’” he said, with solution providers an essential step to establishing controls and guardrails to improve the performance and trustworthiness of AI results.

“A lot of people are starting to believe that, OK, this is the right answer” based on an AI agent’s output to a user’s query, Ziyad said. “And really, it’s about what the AI is paying attention to.”

Phinney’s argument comes as MSPs struggle to turn AI experimentation into measurable operational improvements. Thirty-two percent of MSPs have adopted AI internally, but only 12 percent report measurable productivity gains, according to Nerdio’s 2026 State of the MSP Business Report, released Oct. 20.

The automated end-user computing (EUC) platform provider for Microsoft Cloud technologies surveyed 153 MSP decision-makers responsible for Microsoft 365 and Azure environments and found that just 14 percent operate a revenue-generating AI service line.

Humans Plus AI: A New Model For MSP Work

Humans plus AI could become the most effective workforce teams, Phinney said. MSPs can use large language models to quickly analyze logs, while AI can translate languages, turn prompts into PowerShell scripts and convert files.

Although users shouldn’t let AI agents make decisions for them, they can use agents to evaluate potential decisions, Phinney said. The models produce what he characterized as a weighted average of answers to similar questions because AI is primarily a pattern-detection and generation machine.

“You don’t want this to think for you because it does not know what it’s saying,” he said. “If you’re asking it for advice, you’re going to get consensus advice, and that is as helpful as it is dangerous.”

AI is also enabling faster product building for people with programming expertise and nontechnical users. Phinney used AI to build an agent that delivers a daily briefing drawing from more than 20 newspapers and his personal interests.

Despite not knowing how to use application programming interfaces (APIs), Phinney built another agent that recommends music based on artists he already likes. “This can translate your ideas, your intentions, into code,” he said.

Another agent delivers a project-digest email explaining what Phinney’s team is working on, what employees logged in the business operations platform, where they are stuck, what they completed, what moved during the week and what changed hands. “I don’t have to ask anyone for that information anymore,” he said.

Before job interviews, another agent pulls the candidate’s resume and social profiles and prepares a brief with recommended questions. AI also summarizes hundreds of tickets, call transcripts and messages before Phinney speaks with an upset customer.

His transcription tool turns discovery calls into Markdown files capturing details such as endpoint counts and frustrations with an incumbent MSP, then creates a proposal.

“My close rates have gone way up as a result of this,” he said. “Think about how you can leverage the things that AI is actually good at: attention, translation. There’s a lot you can do with those things, and we’re not doing them yet.”