Sales Discovery Bot

Built by: Meta-Agent Factory
Status: ✅ Unit Tests Passing - Ready for Integration Testing
Session: 2 Complete (2025-06-07)

📚 Documentation

• BUILD_GUIDE.md - Complete build history, session progress, and technical details
• docs/DEVELOPMENT.md - Development setup and local testing
• REPO_LAYOUT.md - Repository structure (auto-generated)

Overview

Type: Conversational Sales Discovery Agent
Framework: LangGraph (stateful conversation management)
Model: GPT-4o
Description: AI-powered sales discovery agent that conducts structured conversations to identify automation opportunities for Insta Agents

Features

• ✅ Structured 6-step conversation flow
• ✅ Automatic MVP identification
• ✅ Partnership tier recommendations ($1,250/$2,500/$5,000)
• ✅ PostgreSQL conversation storage
• ✅ Embeddable in websites and emails
• ✅ Calendly integration for demo booking
• ✅ Redis state management
• ✅ KEDA autoscaling (0-20 pods)
• ✅ Comprehensive test suite

Conversation Flow

1. Business Understanding - 2-3 questions about their business
2. MVP Identification - Focus on one clear automation win
3. Scoping - Get specific details for the MVP
4. Proposal - Formatted MVP proposal
5. Partnership - Recommend appropriate tier
6. Demo Booking - Drive to Calendly link

Quick Start

Local Development

1. Clone and setup:

   git clone https://github.com/Instabidsai/sales-discovery-bot.git
   cd sales-discovery-bot
   python -m venv venv
   .\venv\Scripts\activate  # On Windows
   pip install -r requirements.txt

2. Configure environment:

   cp .env.example .env
   # Edit .env with your keys:
   # - OPENAI_API_KEY
   # - POSTGRES_DSN
   # - REDIS_URL
   # - CALENDLY_URL

3. Start services:

   docker-compose -f docker-compose.dev.yml up -d

4. Initialize database:

   python -m scripts.init_db

5. Run tests:

   pytest tests/ -v

6. Start API:

   uvicorn api.main:app --reload

7. Test the chat:

   curl -X POST http://localhost:8000/chat \
     -H "Content-Type: application/json" \
     -d '{"message": "I run a marketing agency"}'

Deployment

Docker

docker build -t sales-discovery-bot:latest .
docker run -p 8000:8000 sales-discovery-bot

Kubernetes

kubectl apply -f k8s/

Architecture

• Framework: LangGraph v0.0.20 for stateful conversations
• API: FastAPI with health checks and metrics
• Database: PostgreSQL for conversation history
• Cache: Redis for state management
• Scaling: KEDA based on HTTP RPS
• Monitoring: Prometheus metrics + Grafana dashboards

Partnership Tiers

Tier	Monthly Cost	Features
Starter	$1,250	1 AI agent system
Growth	$2,500	Up to 3 concurrent AI systems
Enterprise	$5,000	Unlimited concurrent systems

Session 2 Highlights

• ✅ Fixed Pydantic v2 compatibility
• ✅ Updated to latest LangGraph syntax
• ✅ Fixed datetime deprecation warnings
• ✅ All 12 unit tests passing
• ✅ Local development environment working
• ✅ Ready for integration testing

Next Steps

1. Run integration tests with live services
2. Build and push Docker image to registry
3. Deploy to Kubernetes cluster
4. Test widget embedding
5. Monitor performance metrics

License

Proprietary - Insta Agents

---

Need help? Check BUILD_GUIDE.md for complete technical details and session history.

Generated by Meta-Agent Factory • Last Updated: 2025-06-07