home / mcp / revenue intelligence mcp server

Revenue Intelligence MCP Server

Provides real-time revenue insights by exposing CRM data, lead scoring, churn detection, and model health via MCP interfaces.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "drew6050-revenue-intel-mcp": {
      "command": "python",
      "args": [
        "server.py"
      ]
    }
  }
}

You will run a production-ready MCP server that demonstrates how to integrate real-time ML predictions with revenue-generation workflows. It exposes CRM-like data, scoring and monitoring tools, and prompts that help business teams act on insights while maintaining observability and governance.

How to use

Connect to the Revenue Intelligence MCP Server using an MCP client to access CRM data resources, scoring tools, and model metadata. You can query accounts and leads, run scoring and churn prediction tools, and view feature attributions and drift status. Use the pre-built prompts to analyze accounts, forecast conversion probabilities, and check model health. Ensure you have a running server, then browse or invoke the exposed resources and tools to perform day-to-day revenue intelligence tasks.

How to install

Prerequisites: Python 3.10+ and a Python-friendly environment.

1) Clone or navigate to the project directory named revenue-intel-mcp.

2) Create a virtual environment and activate it.

python -m venv venv
# On macOS/Linux
source venv/bin/activate
# On Windows
venv\Scripts\activate

3) Install dependencies in editable mode to include development tooling.

pip install -e ".[dev]"

4) Run the test suite to verify the installation and environment.

pytest tests/ -v

5) Start the MCP server locally using the provided Python entry point.

python server.py

Configuration and runtime details

Model version: v1.2.3 is recorded in the configuration and metadata for traceability. Lead tier thresholds are Hot (≥70), Warm (40–70), Cold (<40). Feature weights include Company size (20%), Engagement (40%), Industry (20%), and Intent (20%). Churn risk thresholds are Critical (≥70), High (50–70), Medium (30–50), Low (<30). These values guide scoring, explainability, and drift monitoring.

Using the HTTP/stdio MCP configuration

This server is exposed via a local stdio MCP interface. Use the following runtime configuration to connect your MCP client. The entry uses a local Python process running server.py.

{
  "type": "stdio",
  "name": "revenue_intel_mcp",
  "command": "python",
  "args": ["server.py"]
}

Available tools

score_lead

Scores a lead based on company attributes and engagement signals, returning a numeric score, a tier label, feature attributions, and an explanation.

get_conversion_insights

Predicts trial-to-paid conversion probability for an account, returning the probability, engagement signals, and recommended actions.

detect_churn_risk

Analyzes account health to identify churn risk, returning a risk score, risk tier, declining signals, and intervention suggestions.

check_model_health

Monitors the ML system health and performance, returning version, uptime, prediction count, drift status, and accuracy.

log_prediction

Manually log a prediction for monitoring, returning a log ID, timestamp, and success status.