PDQ launches MCP server for AI endpoint management

PDQ has launched early access to a Model Context Protocol server for IT teams, linking endpoint management tasks to AI tools including Claude, Microsoft Copilot, and VS Code.

The server lets users investigate device environments, trigger software deployments, generate reports, and build automated workflows from AI interfaces rather than only through PDQ's console.

It is available free during the early-access period to customers across all subscription tiers and is being distributed through native AI connector directories, making it easier for users to find and connect the service within existing tools.

The launch reflects growing interest in Model Context Protocol, or MCP, an open standard designed to let AI assistants interact with software systems through approved actions and data access. In practice, it gives administrators a way to ask an AI system to perform endpoint management tasks using natural language.

Examples include asking an assistant to show devices missing Google Chrome, identify machines that have not checked in during the last 30 days, deploy the latest Chrome update to a specific group, create an interactive Patch Tuesday report, count devices still running Windows 10, or identify devices missing Zoom and deploy it.

Security controls

The MCP server uses existing administrator-defined permissions, and those controls carry through to AI-based interactions. Administrators can decide whether a given user or integration has read-only, reporting, deployment, or administrative access.

Actions initiated through supported AI clients are authenticated and recorded in the audit log, including the user and AI client involved in the request. Administrators can also review active MCP connections and revoke access through the application connections settings.

That approach addresses a central concern for IT teams considering AI-linked administration tools: how to extend access without loosening oversight. By tying MCP access to the same permission structure already used inside PDQ, the feature is positioned as an additional route into existing systems rather than a separate management layer.

Mark Littlefield, vice president of product at PDQ, set out the company's rationale for the launch.

"IT teams shouldn't have to spend their time translating a straightforward goal into the right combination of clicks, filters, and reports," said Littlefield. "PDQ's MCP server gives admins a more flexible way to investigate their environments and take action while keeping permissions, authentication, and auditing firmly under their control."

Another interface

The MCP server does not replace PDQ's main interface for managing devices, deployments, vulnerabilities, and software. Instead, it offers another way for teams to interact with the endpoint environment, especially when investigating a problem, answering an operational question, or folding endpoint management into a broader workflow handled through AI tools.

The distinction matters because many software suppliers are trying to add AI functions without forcing customers to rebuild established processes. In PDQ's case, the MCP server is framed as an additional interface layered on top of existing product controls and APIs.

The company also drew a distinction between MCP and conventional application programming interfaces. APIs generally require a developer or integration specialist to know which endpoint to call and how to structure the request, while MCP is intended to let AI assistants discover and use approved software actions based on a stated user goal.

For customers, that could reduce the manual navigation needed for routine checks and remedial actions. For software vendors, it offers a standardized route into AI assistants without requiring a separate integration for each platform.

Wider push

Claude users can already access the server through Anthropic's connectors directory, while support for other AI marketplaces is planned. PDQ did not give details on post-trial pricing or packaging, saying those would be announced later.

Founded in 2001 and based in Salt Lake City, PDQ sells endpoint management tools for Windows and macOS devices, covering patching, software deployment, and vulnerability management. It says it serves more than 33,000 customers.

The early-access launch places PDQ among a growing group of software providers adapting their products for AI assistant workflows while maintaining a focus on permissions and traceability in systems used by administrators to make changes across fleets of devices.