Iru MCP in Practice: Real-World Workflows for IT Teams

Iru recently hosted a live session on its new MCP integration, and our own Robby Barnes joined the Iru team and another customer to share the workflows we’ve actually been building in production. If you manage an Apple fleet and you’ve been wondering what all the MCP talk is really good for, this is a practical look at what’s possible today.

The full recording is below, and the rest of this post pulls out the parts we think matter most for IT teams.

What MCP Actually Changes

For years, automating across your tools meant learning each product’s API: different authentication, different data formats, different quirks for every system. Stitching two of them together was real engineering work.

The Model Context Protocol (MCP) changes that. Vendors wrap their APIs in a shared protocol, and AI assistants like Claude can interact with them through plain language. With Iru MCP, that means you can ask for device information, blueprints, app and device inventory, or actions like locking and wiping a device, and then chain those steps across other tools. Read a ticket, disable the account, wipe the associated device, close the ticket, post to Slack: one request, multiple systems, no custom integration code.

Setup is straightforward. In Iru, you create an API token, enable MCP on it, and scope its permissions (read-only or specific capabilities like blueprint management). You then paste the MCP configuration into your assistant of choice. It works with Claude, Gemini, ChatGPT, Cursor, or anything else that supports MCP.

What We’re Building With It

We work with several customers running Iru, and being able to generate clear, one-off reports has been genuinely useful. The best way to work with customers is to be completely transparent about what’s happening on their fleet. When we make a recommendation, we want it grounded in data and legible to every stakeholder at the business, not just the technical team.

A few examples from our side of the session:

Reporting that reads your own scripts. One of the quieter superpowers of the Iru MCP connection is that it can read the results of the scripts you’ve deployed. During an onboarding, an engineering team asked which folders besides /Applications contained app files, whether installers or full applications sitting in Downloads, on the Desktop, or anywhere else. We wrote a script to report that out, then generated a clean, branded report from a single prompt, down to “use this font, use this logo.” Work that used to mean building a front end now takes one well-written request.

Cleaning up the fleet during onboarding. When customers come into Iru at whatever stage they’re in, we need a clear picture of what’s actually on their devices. It’s common to find more than one EDR system installed, and they rarely play nicely together. Surfacing that quickly is the first step to cleaning it up.

Hardware inventory and Apple lifecycle planning. Querying inventory is easy, but the real value comes from combining sources. We pull device age from a serial-number lookup and AppleCare status from Apple’s recently released Apple Business API, then have the assistant merge that with Iru MCP data into a single report, sorted by the coverage expiring soonest. For customers on a lifecycle policy, automatic reports that show exactly what hardware exists, how old it is, and its warranty status make it far easier to plan ahead and budget replacements before they become emergencies.

A Couple of Practical Tips

Two things worth knowing before you start:

• Use the MCP configuration, not the plain token. When you create a token in Iru, the API key shown at the top is not the same as the key inside the MCP configuration block. Copy the MCP configuration. Most assistants will take it as-is and put everything where it needs to go.
• Match the model to the task. MCP itself doesn’t cost anything to use; the cost is whatever assistant you connect it to. We typically run these reports in Claude Code, so it’s just token usage. Use a more capable model like Opus for planning something complex, then refine with a lighter model like Sonnet. Reaching for the most expensive model on every small action adds up fast.

The Iru team also hinted at configuration as code coming in the near future, which pairs naturally with their API-first approach. It’s a space worth watching.

Why This Matters

The throughline across every example in the session was the same: less time on laborious, repeated reporting, and more time on the decisions that actually move a business forward. Consolidating data into a clear format means you can answer the “shoulder-tap” questions without handing everyone access to the underlying tools, and you can deliver recommendations your clients can genuinely understand.

As an Iru partner, we’re putting these workflows to work for the businesses we support every day. If you’d like to talk about modern Apple device management and what automation like this could do for your fleet, reach out below.