You've probably noticed the pattern: every week, another MSP vendor announces AI support. Some ship a chatbot, some ship an "AI assistant," and only a few ship something more substantial: a Model Context Protocol (MCP) server that lets your AI actually reach into the tool and do real work.

If you're trying to figure out where your stack sits in that landscape, this post is for you.

What We Mean by "Has an MCP"

Here's a quick working definition: a tool "has an MCP" when there's a way for an MCP-capable AI client (Claude, ChatGPT, Copilot, Cursor, or whichever your team uses) to reach into that tool through the MCP. Sometimes that means the vendor built their own MCP server. Sometimes it means someone else did (like Conduit, which ships MCP integrations for tools whose vendors haven't gotten there yet).

Either way, the end result for you is the same: your AI can log in, look up data, and take action inside that tool without you having to build a custom integration.

The State of MCP Across the MSP Stack

Here's the picture as of Q4 2026. Note that the space is moving fast enough that any list will be out of date within a quarter. By category:

PSAs. The heavyweights are covered. Autotask, ConnectWise PSA, HaloPSA, Atera, Syncro, and SuperOps all have MCP coverage through Conduit. Kaseya BMS is in beta.

RMMs. Also well-covered. ConnectWise Automate, Datto RMM, NinjaOne, and Liongard have live MCP integrations. Kaseya VSA and Immybot are in beta.

Security. This category has the most breadth. Huntress, SentinelOne, RocketCyber, Blumira, and CIPP are live. ThreatLocker, Blackpoint, and Runzero are in beta. Between the SOC platforms and the endpoint tools, most modern MSP security stacks are already reachable through MCP.

Email security. Abnormal Security, KnowBe4, and SpamTitan are live. Proofpoint, Mimecast, Checkpoint Avanan, and Ironscales are in beta.

IT documentation. Both major platforms are live. IT Glue and Hudu both have full MCP integrations.

Networking, monitoring, and incident management. Auvik, Domotz, BetterStack, PagerDuty, and Rootly are live. Azure MCP covers most Azure monitoring cases.

Business systems. This is where MSP owners often forget to look, and where the wins add up fast. QuickBooks Online, Xero, HubSpot CRM, SalesBuildr, PandaDoc, Pax8, and Sherweb all have live MCP integrations. Microsoft 365 is covered through both the M365 plugin and direct Microsoft Graph access.

If you counted along, that makes a lot of the modern MSP stack already reachable.

What This Actually Gets You

Having an MCP for a tool is not the same as having automation running. It's just the door, not the actual workflow. But once the door is open, the workflows that used to require dedicated engineering projects become questions you can ask.

Some examples:

• Autotask has MCP. Your AI can now triage overnight tickets and post a summary to Slack before your team walks in.
• Datto RMM has MCP. Your AI can now surface every endpoint behind on patches across your whole fleet with one question.
• CIPP has MCP. Your AI can now audit MFA coverage across every M365 tenant you manage in seconds.
• QuickBooks has MCP. Your AI can now flag overdue invoices and cross-reference them against contract renewals.

Each of these used to be a custom project. Now they're a prompt.

The Gaps (and What to Do About Them)

No stack is fully covered yet. If you look at your own tools and see a few that aren't on the list above, it's worth knowing what your options are.

Some things to check:

Is your tool in beta with Conduit? Beta means the integration is being tested. If your PSA or RMM is in the beta list, you can probably start using it now, with the understanding that some of the experience may still be rough.

Is your tool on the plugin roadmap? Conduit's plugin library grows regularly. A tool that isn't there today may show up in the next month.

Is there a bigger vendor category you could shift to? Sometimes the answer to "my tool doesn't have MCP" is "the industry leader in that category does." Whether that's a reason to switch depends on your shop.

Do you want a custom plugin built? Conduit's team writes custom plugins for MSPs who need coverage that isn't in the library yet. If you're running something unusual, that's an option.

Why This Matters (Beyond the Obvious)

The MSPs who get real value from AI in the next 12 months aren't going to be the ones with the fanciest strategies. They're going to be the ones who noticed early that most of their stack was already reachable, and started stringing simple workflows across those connections before their competitors did.

If you've been waiting for AI to "get ready" for MSPs, this is what ready looks like. The doors are open, and the tools your team already uses can talk to your AI. The only question left is which workflows you want to run first.

Want to see the full list of what's supported today? Browse the plugin library.