Featured, the AI co-pilot for PR, has launched an MCP (Model Context Protocol) server that connects Claude, Cursor, VS Code, and other AI agents directly to its workflows and opportunity search. It’s a specific answer to a question the industry has circled for the better part of a year: what do AI agents for PR agencies actually do, once the drafting demos are over?
What “AI agents for PR agencies” means right now
The phrase gets used loosely. Most of what passes for an AI agent in PR software today is a chat window layered on top of a media database, or a monitoring tool that summarizes coverage instead of a person doing it by hand. Useful, but not agentic in the sense the term has come to mean: a system that can take an action without a person copying information between two pieces of software that don’t otherwise talk to each other.
Adoption reflects that gap. Muck Rack’s own research puts AI agent use among PR professionals at just 12%, with more than a quarter of the people it surveyed saying they hadn’t even heard the term. Agencies aren’t short on AI tools. They’re short on AI tools their other AI tools can actually reach.
How Featured’s MCP server works
MCP is the open standard Anthropic introduced in 2024 for connecting AI systems to external tools and data. It’s since been adopted by OpenAI, Google, and Microsoft, and was contributed to the Linux Foundation as neutral infrastructure. Featured’s server exposes three pieces of what was previously locked inside its web app: workflows (browsing templates, reading a workflow’s schedule and run history, and, on paid plans, launching or updating one); a single natural-language search that returns matched media queries, podcasts, journalists, byline publications, LinkedIn influencer opportunities, speaking events with open calls for proposals, and awards; and a read on which accounts are connected to an organization.
Access runs through OAuth 2.1. A user signs in once in a browser, and every tool call afterward is scoped to their own account and permissions — no API key to issue, store, or rotate, and no way for a connected agent to do anything inside Featured that the person who connected it couldn’t already do by hand.
Built for how PR agencies work
The distinction matters most at agency scale. A solo practitioner runs one account. An agency runs a roster — client work spread across a dozen browser tabs and a handful of point tools, handled by an account team that’s usually already working inside an AI assistant for everything else: drafting, research, internal notes. Until now, Featured’s opportunity data and workflow controls lived only inside Featured’s own interface, a separate tab from wherever that other work was happening.
PR agencies are already running client work inside AI tools, not just PR software. The MCP server means the same opportunity data and workflows an account team uses inside Featured are now available inside whatever AI agent they’ve already built their day around, without opening another tab for every client.
Read access with the ability to browse templates, check run history, and search opportunities is available on every plan. Launching a new workflow from a template requires a paid plan, consistent with the usage-based pricing that governs the rest of Featured’s product.
What’s next
Featured’s MCP server is live now with setup instructions for Claude Code, Claude Desktop, Cursor, VS Code, and any other MCP-compatible client at featured.com/mcp.
It’s Featured’s first agent-facing surface, not its last. The company plans to extend the co-pilot into narrative planning, messaging, reporting, crisis communications, and investor relations over time, and to bring each into MCP as it does. For an industry still working out what AI agents for PR agencies actually do beyond the demo, that’s Featured’s answer: not a chatbot bolted onto a database, but the product itself, reachable from wherever an agency’s AI agent already lives.
