🤖 ContactCenterAI

Customer service chatbot built with Dialogflow CX and Cloud Functions. Handles order tracking and FAQs with webhook fulfillment.

Features

• Intent recognition and entity extraction
• Order tracking via webhook
• FAQ responses (shipping, returns, refunds)
• Multi-turn conversation flows
• Error handling and logging

Architecture

User → Dialogflow CX → Cloud Functions Webhook → Backend API
                    ↓
            Dynamic Responses

Tech Stack

• Dialogflow CX - Intent matching, NLU
• Cloud Functions - Python webhook
• pytest - Testing

Setup

1. Clone the repo:

git clone https://github.com/anushreejha/ContactCenterAI.git
cd ContactCenterAI

2. Configure GCP:

export PROJECT_ID="your-project-id"
gcloud config set project $PROJECT_ID
gcloud services enable dialogflow.googleapis.com cloudfunctions.googleapis.com

3. Deploy webhook:

cd cloudfunctions/webhook
gcloud functions deploy dialogflow-webhook --runtime python311 --trigger-http --entry-point webhook --region us-central1

4. Import intents from dialogflow/intents/ using the Dialogflow CX console at dialogflow.cloud.google.com/cx

Project Structure

ContactCenterAI/
├── cloudfunctions/
│   └── webhook/
│       ├── main.py
│       └── requirements.txt
├── dialogflow/
│   ├── intents/
│   │   ├── order_tracking.json
│   │   ├── faq_shipping.json
│   │   ├── faq_returns.json
│   │   └── welcome.json
│   └── entities/
│       └── order_id.json
├── tests/
│   ├── __init__.py
│   ├── test_webhook.py
│   └── fixtures/
│       └── sample_requests.json
├── docs/
    ├── ARCHITECTURE.md
│   └── sample_conversation.md
├── scripts/
│   ├── deploy.sh
│   └── setup_gcp.sh
├── .gcloudignore
├── .gitignore
├── README.md
└── requirements-dev.txt

Usage

Order Tracking: User: What's my order status? Bot: Please provide your order ID. User: ORD-12345 Bot: Your order is in transit. Expected delivery: Dec 28, 2025.

FAQ: User: What's your return policy? Bot: We offer 30-day returns for unused items.

Testing

Run tests:

pytest tests/test_webhook.py -v

Test locally:

cd cloudfunctions/webhook
functions-framework --target=webhook --debug

Configuration

Local dev environment variables (.env):

PROJECT_ID=your-project-id
ORDER_API_URL=<api-link>

Future Enhancements

• Multi-language support (Hindi)
• Integration with live order API
• Sentiment analysis for escalation priority
• Chat history persistence