PDQ MCP SERVER
Automate IT work within your own AI tools
Connect your own AI tools like Claude Code, OpenAI Codex, and Microsoft Copilot to build workflows, reports, and automations. Start with simple prompts, then move on to advanced multistep instructions and repeatable skills:
“How many devices are still running Windows 10?”
“Create an automation to keep every device up –to date with the latest version of Zoom”
“Build a custom Patch Tuesday impact report”
“Create a dashboard of devices without Firefox or Chrome in PDQ Connect that I can share with my IT Director”
“If I have to do this twice, it should be automated”
A SYSADMIN, PROBABLY
Or, just ask PDQ
Skip digging into logs until you’re cross-eyed or doing the diagnostic legwork between spotting a problem and understanding what’s going on. PDQ’s AI assistant can investigate using the context already in PDQ, not just what you type into a chat box. It can break down what happened during a deployment, connect the results to relevant device data, and propose targeted diagnostics to gather additional evidence when an issue needs a closer look.
PDQ AI is observable and configurable by design
Comprehensive audit logs
Every action taken with PDQ AI assistants and through the PDQ MCP server is logged — so you get a complete history of what was done, by whom, and from where.
No customer data training
We have contractual agreements with our AI subprocessors that prohibit the use of customer data to train their models.
Organization-wide AI controls
Choose whether to enable AI features across your organization so you can adopt AI capabilities when it makes sense for your organization.
Role-based access
PDQ AI capabilities mirror individual permissions and privileges within the PDQ console — so you can deploy and control privileges simply.
Secure encryption
When you use AI capabilities at PDQ, your data is encrypted in transit using TLS 1.2 or greater.
FAQs
What AI features does PDQ offer?
PDQ currently offers two opt-in ways to use AI in PDQ.
The PDQ AI assistant is built into PDQ and can help build packages, understand deployment results, and investigate device issues using supported PDQ context and tools.
The PDQ MCP server lets compatible external AI tools work with supported PDQ data and capabilities.
What can I ask the PDQ AI assistant?
The assistant currently works with packages, deployments, and devices. For example:
“Build a package to deploy [app].”
“What actually happened during this deployment?”
“Why did this deployment fail on this device?”
“What's filling up this drive?”
How is the PDQ AI assistant different from pasting something into a chatbot?
A general-purpose chatbot works with the context you give it. The PDQ AI assistant can work with supported context already in PDQ, such as package details, actual deployment output, and device information.
That means you don't have to gather and paste all of that information into another tool before you can ask a question about it.
Can the AI assistant run commands, scripts, or deployments on my devices? Does it automatically fix problems?
No. Any action, whether it’s running a command, creating a package, deploying a package, patching a fix, running a script, etc., requires a human to approve it before it happens. You stay in control at all times.
What is the PDQ MCP server?
MCP, short for Model Context Protocol, is the open standard that makes this possible: a secure way for an AI assistant to reach into a tool like PDQ and act on your behalf, using only the access your account already has.
The PDQ MCP connects your AI tool directly to PDQ so you can ask questions and take action by describing what you want to do in plain English.
The available tools cover searching devices, groups, packages, deployments, automations, and vulnerabilities in your fleet, along with the ability to deploy packages, create groups, and set up automations. The assistant inherits your existing role and permissions exactly as they work in the console — if you can’t do something there, you can’t do it through MCP either.
Does PDQ AI cost anything to use?
No, PDQ AI is free to use for all customers during the early access period that will extend at least through December 31, 2026.
Does PDQ train AI using customer data?
By default, PDQ and its AI subprocessors do not use customer data to train any models. We have contractual agreements in place with our AI subprocessors that specifically prohibit the use of customer data to train their models.
How is PDQ AI different from a foundation model or self-built agent?
PDQ AI includes both:
(1) Our own AI assistant features, which are built on top of foundation models with specific PDQ context, tools, and instructions and adapt to individual user prompts.
(2) Our AI MCP server, which enables your own agentic AI tools to access PDQ tools — meaning any model-specific behavior is determined by an individual user’s tools, prompts, and context.
Can I turn PDQ AI off for my organization?
Yes, 100% of AI features can be disabled in the Settings AI Tools menu. This is an org-wide setting only. Enabling/disabling features are captured in the audit logs.