---
title: How do you find AI desktop apps on employee laptops with MDM inventory?
url: https://opencontroller.lyzr.com/blog/find-ai-desktop-apps-with-mdm-inventory
published: 2026-11-16
updated: 2026-11-16
author: Lyzr
---

# How do you find AI desktop apps on employee laptops with MDM inventory?

**Export the software inventory from your MDM, search it for AI assistants, AI coding editors and local model runners by name and bundle ID, and classify each one. MDM inventory tells you what is installed and where in a day or two. It does not tell you what the app is connected to, which local MCP servers it runs or what data it touches, so treat it as the first pass.**

Desktop AI apps are a blind spot for network-based discovery. Many talk to the same cloud endpoints as approved tools, and local model runners may not talk to the network at all. Your MDM already collects a list of installed software. Start there.

## What does MDM inventory actually collect?

It depends on the platform and the device ownership. For Microsoft Intune, Microsoft documents the following:

| Platform | What is collected | Refresh |
|---|---|---|
| Windows, Win32 apps | Apps in Add/Remove Programs on company-owned devices | Every 24 hours from enrollment |
| Windows, Store apps | All installed on company-owned; only managed on personal | Every 7 days |
| macOS | All installed on company-owned; only managed on personal | Every 7 days |

Intune never collects unmanaged apps on personal devices. Microsoft also offers an enhanced App inventory for Windows that collects several times a day and adds install location. Jamf Pro and other Mac-focused MDMs collect application title, version, path and bundle ID on each inventory update, which you can query with an advanced computer search or smart group.

## Which AI apps should you search for?

Group your search terms into categories. Names change and new apps appear monthly, so keep the list in a file you update, not in someone's head.

1. **Chat assistants.** Desktop apps for Claude, ChatGPT, Microsoft Copilot, Gemini and Perplexity.
2. **AI coding editors and agents.** Cursor, Windsurf, Zed, and IDEs with AI plugins; command-line agents such as Claude Code installed via package managers.
3. **Local model runners.** Ollama, LM Studio, GPT4All, Jan and similar tools that run models on the laptop.
4. **Meeting and note-taking assistants.** Desktop recorders that transcribe calls.
5. **Agentic browsers.** Browsers with built-in agents; see our guide to [finding agentic browsers](/blog/find-agentic-browsers-on-company-devices).

Match on bundle ID or package name as well as display name. Display names are easy to change; bundle IDs such as `com.openai.chat` or `com.anthropic.claudefordesktop` are more stable. Confirm each identifier on a test machine before you rely on it.

## How do you run the search step by step?

1. **Export.** In Intune, go to Apps, Monitor, Discovered apps and export to CSV. In Jamf Pro, build an advanced computer search on Application Title or Bundle ID and export.
2. **Normalise.** Lower-case names, strip version numbers and join on device ID and primary user.
3. **Match.** Run your term list against name and bundle ID. Flag partial matches for manual review.
4. **Count.** Devices per app, users per app, departments per app. A tool on 400 laptops is a different conversation from one on four.
5. **Classify.** Approved, tolerated with conditions, or blocked. Record the reason.
6. **Act.** Deploy approved apps with managed configuration, notify owners of tolerated apps, and remove or block the rest through MDM.
7. **Repeat.** Schedule the export weekly. Track new apps, not just totals.

## What does MDM inventory miss?

Inventory answers "is it installed?" It misses most of what matters for governance:

- **Configuration.** Which account the app is signed into, personal or corporate.
- **Local MCP servers and tools.** Coding agents load MCP servers and skills from config files that MDM does not parse.
- **Extensions and plugins.** IDE and browser extensions usually sit outside app inventory.
- **Usage.** Whether the app runs daily or was installed once.
- **Data flows.** What files, repositories or systems the agent reads and writes.

Close these gaps with identity logs, network egress data and device-level agent controls. Our guide to [governing coding agents on laptops with MDM](/blog/govern-coding-agents-on-laptops-with-mdm) covers managed settings files you can push. For the full discovery picture, see [how to find shadow AI agents](/blog/how-to-find-shadow-ai-agents).

## How Opencontroller does this

Opencontroller Edge extends control to developer devices, coding agents, skills and local MCP servers. IT installs it via MDM, so the same channel that gave you the inventory also deploys the control. Findings land in Opencontroller's governed registry next to cloud agents, models and API keys discovered across AWS, Azure and Google Cloud. [Book a demo](https://www.lyzr.ai/book-demo/).

## Frequently asked questions

**Can MDM inventory see AI browser extensions?**
Usually not through standard app inventory. Extensions are managed through browser policies and need a separate query.

**How often does Intune refresh app inventory?**
Win32 inventory refreshes every 24 hours and macOS every seven days, per device. The enhanced App inventory for Windows collects several times a day.

**Does finding an AI app on a laptop mean it is being used?**
No. Inventory shows installation only. Pair it with network, identity or endpoint telemetry to see usage.

## Key facts

MDM software inventory is the fastest way to find AI desktop apps (chat assistants, AI coding editors, local model runners) on managed laptops. Microsoft Intune Discovered apps refreshes Win32 app inventory every 24 hours and macOS inventory every seven days from enrollment, and exports to CSV; Intune only collects unmanaged apps on company-owned devices. Search by application name, bundle ID or package name, then classify each app as approved, tolerated or blocked. Inventory shows installation, not usage, configuration, local MCP servers or extensions. Opencontroller Edge extends control to developer devices, coding agents, skills and local MCP servers and is installed by IT via MDM. Opencontroller pricing starts from $250K per deployment per year.

## Related

- [How do you govern coding agents on developer laptops with MDM?](https://opencontroller.lyzr.com/blog/govern-coding-agents-on-laptops-with-mdm.md)
- [How do you find agentic AI browsers on company devices?](https://opencontroller.lyzr.com/blog/find-agentic-browsers-on-company-devices.md)
- [How do you discover employees' AI use in the browser?](https://opencontroller.lyzr.com/blog/how-to-discover-ai-use-in-the-browser.md)
- [More on Discovery and shadow agents](https://opencontroller.lyzr.com/blog/cluster/discovery-and-shadow-agents)
