MCP Server

We provide an MCP server to help you leverage LLMs to explore workforce data and insights. This page describes the overall capabilities and what you need to know to use the MCP with your LLM of choice.

What is MCP?

Model Context Protocol, MCP, is a technology that allows any chatbot (Claude, ChatGPT, etc) to interact with a backend data provider, like Live Data/Workforce. It's a "connector" that instructs the chatbots how to get and analyze data from our system.

Today, support for MCP varies across chatbots, but over time this support is likely to become more standardized across the industry. If you are familiar with MCPs or willing to get a little more technical, connecting our MCP to your chatbot will allow you to have rich conversations about workforce data while in your native chat interface.

Client Support

Every chatbot "client" has its own stated support for MCP servers and the various standards that are currently being matured. As such, the Workforce.ai MCP server may not work with all clients. This list will be updated over time as more experimentation is done.

Supported Clients

Claude.ai (setup instructions below)
Claude Desktop (setup instructions below)

Unsupported Clients

ChatGPT * coming soon!
Cursor

MCP Server URL

The workforce.ai MCP server is a hosted server. It can be added to your tool of choice by providing the MCP remote server URL, which is https://mcp.workforce.ai. Refer to your LLM documentation for details on how to add this server.

Authentication

The workforce.ai MCP server supports OAuth authentication as specified in the MCP spec. After adding the MCP server to your LLM, you should be redirected to an authentication page within workforce.ai where you can approve access to your workforce.ai account. If you don't have a workforce.ai account, you can create one as part of this process.

Capabilities

The MCP server is capable of answering a variety questions about the workforce -- these questions generally are centered around either companies or people.

Company Analysis

The MCP can answer questions related to the workforce of a given company or a set of companies. These could be used to dive deep into a specific company to understand workforce dynamics for hiring or investment purposes, or could be used to compare companies for benchmarking or other purposes.

Some examples of the types of company-focused questions are:

How many ppl currently at OpenAi previously worked at Goldman Sachs?
Which faang has the shortest average employee tenure?
Which ibanks have the highest percentage of employees with 5+ years tenure?
What is the dominant location of the sales teams at Oracle vs. Workday
What percentage of consultants at McKinsey have MBAs vs. undergrad only?

Person Discovery & Analysis

The MCP can also help identify people or groups of people based on your query. You can find specific people for recruiting purposes or get answers on characteristics of groups of people.

Some examples of people-focused questions would be:

Look at a selection of companies of your choice. How long do people typically stay in their first job out of college?
Find all Stanford CS graduates that work at early-stage AI companies.
Find CTOs who previously worked at both Microsoft and Amazon
Find all former Tesla employees now working at Rivian or Lucid Motors

MCP Client Setup

Claude (web and desktop)

In order to add the Workforce.ai MCP to Claude, you must have a paid plan with Anthropic.

Open Claude and click the Search and Tools icon, and then Manage Connectors
Press Add Custom Connector
For the Name, add a description name, such as Workforce AI
For the Remote MCP server URL, enter https://mcp.workforce.ai
Press Add
Once you see Workforce AI in the list, click Connect
This will take you to workforce.ai to authorize Claude to connect to your workforce.ai account. Approve the request
Now when asking Claude questions, prompt it to use Workforce, such as "using workforce ai, how many engineers work at meta"

Support

For any questions or issues, contact us at [email protected]