# Polygate MCP Server

A Model Context Protocol server that exposes Polygate business tools to AI clients through a standardized interface.

## What It Does

The server allows an MCP-compatible AI client to discover and call business tools such as:

- finding inactive customers
- searching internal knowledge
- drafting customer follow-ups
- creating sales tasks

This demonstrates how an AI assistant can move beyond answering questions and interact directly with business systems.

## Why MCP Matters

Businesses often have valuable capabilities spread across internal APIs, databases, CRM systems, documents, and operational software.

Without a common interface, every AI integration requires custom work.

MCP provides a standardized way for AI systems to access business tools and data.

The Polygate MCP Server demonstrates this pattern:

AI Client  
↓  
Model Context Protocol  
↓  
Polygate MCP Server  
↓  
Business Tools / Data / Workflows

## Current MCP Tools

### `get_inactive_customers`

Finds customers that have been inactive beyond a specified number of days.

Example use:

> Find every customer we haven't contacted in 90 days.

### `search_polygate_knowledge`

Searches Polygate's internal business knowledge.

Example use:

> What does Polygate Lead Recovery do?

### `create_customer_followup`

Generates a follow-up draft for an existing customer.

Example use:

> Prepare a follow-up for Prime Suites.

### `create_sales_task`

Creates and assigns a business task.

Example use:

> Create a task for sales to follow up Prime Suites tomorrow.

## Business Applications

This architecture can be expanded into an enterprise MCP layer connecting AI assistants to:

- CRM systems
- customer databases
- ERP platforms
- internal knowledge bases
- hotel management systems
- task management software
- email systems
- messaging systems
- operational databases
- government systems
- internal APIs

## Example Enterprise Workflow

A user could ask:

> Find every dormant hospitality customer, prepare a reactivation message, and create tasks for the sales team.

An MCP-enabled AI agent could:

1. query customer records
2. identify inactive accounts
3. retrieve relevant customer context
4. draft the outreach
5. create follow-up tasks
6. return the result to the user

This turns natural language into business action.

## Tech Stack

- Python
- Model Context Protocol SDK
- MCPServer
- Pytest
- JSON sample data

## Project Structure

```text
polygate-mcp-server/
├── app/
│   ├── __init__.py
│   ├── server.py
│   └── tools.py
├── data/
│   ├── customers.json
│   └── knowledge.json
├── tests/
│   └── test_tools.py
├── .gitignore
├── requirements.txt
└── README.md