Stop Calling It AI Strategy

Every MSP and their dog now has an "AI strategy." I've sat through the vendor pitches. I've read the whitepapers. I've watched the webinars where someone with "AI Evangelist" in their title explains how AI will revolutionize managed services.

Here's what most "AI strategies" actually are: a ChatGPT Enterprise subscription, a vague plan to "leverage AI for ticket automation," and a PowerPoint deck that gets shown to customers during QBRs.

That's not strategy. That's tourism.

The gap between using AI and building with AI

There's a meaningful difference between using AI tools and building AI into your operations. Using Copilot to write emails faster is fine. It's productivity tooling. But it's not a strategic advantage because your competitor can do the same thing by tomorrow.

Strategic AI integration means building systems where AI is embedded in your workflows — where it's doing work that creates compounding value over time. That means:

• Custom models trained on your data: Your ticket history, your resolution patterns, your customer environments. This is your moat.
• Agent systems that automate multi-step processes: Not just summarizing a ticket, but triaging it, pulling relevant KB articles, suggesting a resolution, and escalating with context when it can't resolve automatically.
• Evaluation frameworks that improve over time: Every interaction makes the system better. Every failure mode gets captured and addressed.

Why most MSPs won't do this

Because it requires engineering, not procurement. You can't buy this off the shelf. You can buy components — LLM APIs, vector databases, orchestration frameworks — but the integration, the customization to your specific workflows, the evaluation and iteration? That's engineering work.

Most MSPs don't have engineering teams. They have operations teams that configure vendor tools. And there's nothing wrong with that for most of the business. But if your "AI strategy" is waiting for ConnectWise or Datto to ship AI features, you're outsourcing your competitive advantage to your PSA vendor's product roadmap.

What to actually do

Start small and build internal capability:

Pick one workflow that's high-volume and well-understood. Ticket triage is the classic example.

Build the data pipeline first. Get your ticket data into a format you can work with. Clean it. Label it. Understand it.

Build a simple prototype. Use Claude or GPT-5 with your data as context. Measure how well it performs against your current process.

Iterate based on real metrics. Not "it seems to work," but actual precision/recall numbers on a test set.

This is slower than buying a vendor solution. It's also the only approach that builds a real competitive advantage. The MSPs that invest in this now — in the actual engineering, not the slide decks — will be operating at a fundamentally different level in two years.

Stop calling it AI strategy. Start calling it engineering. Then do it.