Introducing Falcon MCP: Your LLM, Falcon's Data

by Yubin Park, PhD, Co-Founder

Healthcare data analysis works best when you can stay in flow. You start with a question, get an answer, and immediately ask the next one—without switching tools, re-uploading files, or re-explaining context. We've been working on something that makes this possible across platforms: Falcon MCP.

What Is Falcon MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants connect to external data sources and tools. With the Falcon MCP endpoint, your enterprise LLM platform—whether that's Claude Desktop, the OpenAI API, or your own internal AI tooling—can reach directly into Falcon's proprietary healthcare data.

That means the AI you already use every day can answer healthcare-specific questions that would otherwise require logging into a separate system.

Workflow showing how an enterprise LLM connects to Falcon data and hands off to Falcon AI for deeper investigation

How It Works

The workflow has two parts.

Query Falcon data from your LLM. Ask your AI assistant a question—say, the top billers for a procedure in 2025. Behind the scenes, the assistant calls the Falcon MCP endpoint, retrieves the answer from Falcon's data, and surfaces it directly in your conversation. No export, no copy-paste, no context switch.

Continue in Falcon when you need to go deeper. When a quick answer turns into a full investigation, you can hand off the conversation to Falcon AI with a single click. Your context—the query, the findings, the thread of reasoning—carries over completely. Falcon's advanced analytics and specialized investigation tools are waiting on the other side, and you pick up exactly where you left off.

Who This Is For

Falcon MCP is available to enterprise organizations that have:

  • An active Enterprise agreement with OpenAI or Anthropic
  • A BAA (Business Associate Agreement) in place with their LLM provider

These requirements aren't bureaucratic hurdles—they're the baseline that makes it safe to move patient-adjacent data between platforms. We take the compliance side seriously, and we'll only connect through channels where both sides have the right protections in place.

Where We Are Today

We're currently testing with a select group of clients. Early results confirm what we expected: when organizations can bring their own LLM workflows to Falcon's data, the questions get better and the investigations move faster. The friction of switching tools disappears, and analysts can follow a line of inquiry without losing momentum.

We're looking to expand. As more healthcare organizations stand up their own LLM platforms—internal deployments, cloud-hosted enterprise accounts, custom-built assistants—the opportunity to connect them securely to specialized data like Falcon's grows with it. If your organization has an enterprise LLM setup and a signed BAA, we want to talk.

What's Next

MCP is still early, and the ecosystem is moving fast. We're watching which platforms organizations are actually building on and prioritizing connections accordingly. We're also working on making the context handoff even richer—so when you continue a conversation in Falcon, it's not just the text that carries over, but the analytical state behind it.

If you're interested in piloting Falcon MCP with your organization, reach out to us at sales@falconhealth.ai. We'll walk you through the technical requirements and get you connected.

More articles

Research Brief: Anomalous Urinary Catheter Billing and the ACO Exposure

The first in our research brief series. Medicare spending on urinary catheter codes A4352 and A4353 jumped roughly 20x in two years — and ACOs absorbed the cost. Download the full white paper.

Read more

How ACOs Should Approach FWA Campaigns: The Case for Structured Outreach

Most ACOs sit on claims data that could surface millions in billing irregularities — but lack a clear strategy for acting on it. Here's how structured, evidence-based outreach gets providers to self-correct without formal enforcement.

Read more

Ready to put an AI co-analyst to work for your team? Let's talk.

Our offices

  • Palo Alto, CA
    Corporate HQ
  • Atlanta, GA
    Technology Branch