Parlant Guidelines vs Traditional LLM Prompt: Life Insurance Agent Demo

This project demonstrates the advantages of Parlant's structured approach over traditional monolithic LLM prompts for building conversational agents.

Quick Start

Terminal 1 - Start the server:

uv run parlant_agent_server.py

Terminal 2 - Run the comparison:

uv run demo_comparison.py

Demo Queries

The demo tests 5 realistic scenarios:

• Policy replacement with critical warnings
• Coverage calculation with specific parameters
• Health condition impact assessment
• Mixed topics with boundary maintenance
• Decision making with conflicting rules

Project Structure

parlant-conversational-agent/
├── parlant_agent_server.py      # Parlant agent with tools & guidelines
├── demo_comparison.py            # Main comparison demo runner
├── traditional_llm_prompt.py     # Monolithic prompt approach
├── parlant_client_utils.py      # Parlant API client utilities
├── rich_table_formatter.py      # Beautiful console table rendering
└── pyproject.toml               # Project dependencies (uv)

Setup

uv sync  # Install dependencies

Requirements

• Python 3.10+ (required for Parlant)
• uv package manager
• OpenAI API key in .env file

📬 Stay Updated with Our Newsletter!

Get a FREE Data Science eBook 📖 with 150+ essential lessons in Data Science when you subscribe to our newsletter! Stay in the loop with the latest tutorials, insights, and exclusive resources. Subscribe now!

---

Contribution

Contributions are welcome! Please fork the repository and submit a pull request with your improvements.