🤖 BandhanAI - AI-Powered Customer Relationship Management Agent

BandhanAI is an intelligent customer service and marketing automation agent designed specifically for Indian e-commerce platforms. Built with LangGraph and LangChain, it leverages AI to understand customer behavior, create targeted marketing campaigns, and automate customer communications through email and Slack integrations.

✨ Features

🎯 Smart Customer Analytics

• Customer Segmentation: Automatically categorize customers based on behavior, purchase history, and engagement
• Churn Risk Analysis: Identify at-risk customers and proactively engage them
• Purchase Pattern Recognition: Understand customer preferences and buying cycles

📧 Advanced Email Marketing

• Personalized Campaigns: Create highly targeted email campaigns based on customer data
• HTML Email Templates: Send beautifully formatted HTML emails
• Campaign Types: Support for 9 different campaign types (loyalty, referral, re-engagement, etc.)
• Gmail Integration: Seamless email sending through MCP Gmail server (Pinedream)

💬 Slack Integration

• Real-time Updates: Share campaign status and insights with your team
• Error Reporting: Automatic error notifications and issue reporting
• Success Celebrations: Share milestones and achievements

🔄 Intelligent Automation

• LangGraph Workflows: Advanced AI agent workflows with human-in-the-loop capabilities
• Database Integration: Direct connection to Supabase PostgreSQL for real-time data access
• MCP Architecture: Modular Component Protocol for extensible integrations

🏗️ Architecture

BandhanAI/
├── 📁 .langgraph_api/          # LangGraph API configurations
├── 📄 config.py               # Environment and path configuration
├── 📄 main.py                 # Main application entry point
├── 📄 prompts.py              # AI agent system prompts
├── 📄 server.py               # MCP marketing server
├── 📄 graph.py                # LangGraph workflow definitions
├── 📄 frontend.py             # Web interface (if applicable)
├── 📄 example_mcp_config.json # MCP server configuration template
├── 📄 langgraph.json          # LangGraph project configuration
└── 📄 requirements.txt        # Python dependencies

🚀 Quick Start

Prerequisites

• Python 3.13 or higher
• PostgreSQL database (Supabase recommended)
• Gmail account for email sending
• Slack workspace for notifications
• Google Gemini API key

Installation

1. Clone the repository

   git clone https://github.com/MohammedMusharraf11/BandhanAI.git
   cd BandhanAI

2. Install dependencies

   pip install -r requirements.txt

3. Environment Setup

   Create a .env file in the root directory:

   # Database Configuration
   SUPABASE_URI=postgresql://your_username:your_password@your_host:5432/your_database
   
   # Google Gemini API
   GOOGLE_API_KEY=your_gemini_api_key
   
   # Slack Integration
   SLACK_BOT_TOKEN=xoxb-your-slack-bot-token
   SLACK_CHANNEL_ID=your_channel_id
   
   

4. Database Setup

   Ensure your PostgreSQL database has the required tables:

   -- Customer data table
   CREATE TABLE public.crm (
     customer_id bigint NOT NULL,
     name text NOT NULL,
     email text NOT NULL,
     region text,
     age bigint,
     income bigint,
     segment text,
     last_purchase timestamp without time zone,
     total_spend double precision,
     product_category text,
     churn_risk double precision,
     feedback_score double precision,
     products text[],
     CONSTRAINT crm_pkey PRIMARY KEY (customer_id)
   );
   
   -- Marketing campaigns table
   CREATE TABLE public.marketing_campaigns (
     id uuid NOT NULL DEFAULT gen_random_uuid(),
     name text NOT NULL,
     type text CHECK (type = ANY (ARRAY['loyalty'::text, 'referral'::text, 're-engagement'::text, 'at risk'::text, 'new customer'::text, 'champion'::text, 'about to sleep'::text, 'lost'::text, 'potential loyalist'::text])),
     description text,
     status text DEFAULT 'draft'::text,
     created_at timestamp without time zone DEFAULT now(),
     CONSTRAINT marketing_campaigns_pkey PRIMARY KEY (id)
   );
   
   -- Campaign emails table
   CREATE TABLE public.campaigning_emails (
     id bigserial NOT NULL,
     campaign_id uuid NOT NULL REFERENCES public.marketing_campaigns(id),
     customer_id bigint NOT NULL,
     email text NOT NULL,
     subject text NOT NULL,
     body text NOT NULL,
     sent_at timestamp without time zone DEFAULT now(),
     status text DEFAULT 'sent'::text,
     opened boolean NOT NULL DEFAULT false,
     CONSTRAINT campaigning_emails_pkey PRIMARY KEY (id)
   );

5. Configure MCP Servers

   Copy example_mcp_config.json to mcp_config.json and update with your settings:

   {
    "mcpServers": {
   
   
      "postgres": {
        "command": "npx",
        "args": [
          "-y",
          "@modelcontextprotocol/server-postgres",
          "${SUPABASE_URI}"
        ],
        "transport": "stdio"
      },
   
   
      "marketing": {
        "command": "python",
        "args": [
            "server.py"
        ],
        "transport": "stdio"
      },
   
      
      "slack": {
        "command": "npx",
        "args": [
          "-y",
          "@modelcontextprotocol/server-slack"
        ],
        "env": {
          "SLACK_BOT_TOKEN": "${SLACK_BOT_TOKEN}",
          "SLACK_TEAM_ID": "${SLACK_TEAM_ID}",
          "SLACK_CHANNEL_IDS": "CHANNEL-ID"
        },
        "transport": "stdio"
      },
   
      "pd": {
      "command": "npx",
      "args": [
        "-y",
        "supergateway",
        "--sse"
        
      ],
      "transport": "sse",
      "url": "https://mcp.pipedream.net/API-KEY/gmail"
   
    }
    }

Running the Application

1. Start the main application

   python main.py

2. Interact with Ralph (the AI agent)

   ---- 🤖 Assistant ----
   
   Hello! I'm Ralph, your AI-powered customer service and marketing automation agent for BandhanAI...
   
   User: Analyze our high-value customers and create a loyalty campaign
   

🎯 Campaign Types

BandhanAI supports 9 different marketing campaign types:

Campaign Type	Description	Target Audience
Loyalty	Thank high-value customers	Long-term, high-spending customers
Referral	Encourage customer referrals	Satisfied, engaged customers
Re-engagement	Win back inactive customers	Customers with no recent purchases
At Risk	Prevent customer churn	High churn risk score customers
New Customer	Welcome and onboard	Recently acquired customers
Champion	Reward best customers	Top-tier customers
About to Sleep	Re-activate before churn	Declining engagement customers
Lost	Win back churned customers	Churned customers
Potential Loyalist	Nurture promising customers	New customers with high potential

💡 Usage Examples

Creating a Customer Segment Analysis

User: Analyze our customers who haven't purchased in the last 3 months and have a high churn risk

Launching a Re-engagement Campaign

User: Create a re-engagement campaign for customers in the Mumbai region who haven't bought anything in 60 days

Sending Personalized Emails

User: Send a loyalty email to our top 10 customers thanking them for their continued support

Slack Reporting

User: Post a summary of today's campaign performance to our marketing Slack channel

🛠️ Technology Stack

• AI Framework: LangGraph + LangChain for intelligent workflows
• LLM: Google Gemini for natural language processing
• Database: Supabase (PostgreSQL) for customer data
• Email Service: Gmail via MCP (Pinedream)
• Communication: Slack API for team notifications
• Architecture: MCP (Model Context Protocol) for modular integrations

🔧 Configuration

MCP Server Configuration

The application uses MCP servers for external integrations:

• Marketing Server: Handles campaign creation and email sending
• Gmail Server: Manages email delivery through Pinedream MCP
• Slack Server: Handles team notifications and updates

Agent Configuration

Ralph (the AI agent) can be configured through prompts.py:

• System Prompt: Defines the agent's personality and capabilities
• Database Schema: Provides context about available data
• Campaign Types: Defines available marketing campaign types
• Email Templates: Configures email formatting and personalization

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines for details.

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Inspired by crm-agent by Kenneth Liao
• Built with LangGraph and LangChain
• Email integration powered by Pinedream MCP

📞 Support

If you encounter any issues or have questions:

1. Check the Issues page
2. Create a new issue with detailed information
3. Join our community discussions

---

Built with ❤️ for the Indian e-commerce ecosystem