Agent Purpose

Transform AI pricing from guesswork to science with ROI-based optimization.

For AI companies, traditional pricing models fail to capture your technology's true value. This intelligent agent analyzes your pricing ecosystem to deliver:

• Segment-specific ROI analysis for optimal pricing per customer type
• Dynamic recommendations based on usage patterns and value capture
• A/B testing frameworks for experimental validation
• Data-driven strategies that maximize revenue and reduce churn

Essential Documentation Required

📋 Product Docs

• ICP descriptions with use cases
• Feature catalogs with value propositions
• Customer success stories with ROI metrics

🏗️ Infrastructure Docs

• Usage metrics and cost structures
• Scalability constraints and benchmarks
• Integration requirements

🤖 AI Usage Docs

• Model performance metrics
• Token consumption patterns
• Usage analytics and adoption data

Ready to optimize your AI pricing? Follow the setup instructions below.

Agent Architecture

This diagram illustrates the complete workflow of the AI Pricing Research Consultant, showing how data flows through various specialized agents to deliver comprehensive pricing recommendations.

The architecture follows a modular design with clear separation of concerns:

• Data ingestion and storage layers
• Specialized AI agents for different analysis phases
• Orchestration layer coordinating the workflow
• Multiple AI model integrations for optimal performance

Each agent is designed to handle specific aspects of pricing analysis, from initial product offering assessment to experimental recommendation generation.

---
config:
  theme: mc
  look: classic
  layout: dagre
---
flowchart TD
    A["CLI Entry Point<br>main.py"] --> B{"Command Type"}
    B -- "--create" --> C["Data Creation Flow"]
    B -- "--orchestrator" --> D["Orchestrator Flow"]
    B -- "--delete" --> E["Delete Operations"]
    B -- "--list" --> F["List Operations"]
    C --> C1["JSON Input"]
    C1 --> C2["connectors.py<br>create_from_json_file"]
    C2 --> C3["MongoDB<br>Product, PricingModel,<br>CustomerSegment"]
    D --> D1["orchestrator.py<br>final_agent"]
    D1 --> D2["Product Offering Agent<br>deepresearch/product_offering.py"]
    D2 --> D3["Segmentwise ROI Agent<br>deepresearch/segmentwise_roi.py"] & AI1["OpenAI o3-deep-research<br>with web search, file search,<br>code interpreter"] & DB1[("MongoDB<br>Product Data")]
    D3 --> D4["Pricing Analysis Agent<br>deepresearch/pricing_analysis.py"] & AI2["OpenAI GPT-5<br>with reasoning"] & DB2[("MongoDB<br>CustomerSegment<br>CustomerUsageAnalysis")]
    D4 --> D5["Value Capture Analysis Agent<br>deepresearch/value_capture_analysis.py"] & AI3["OpenAI GPT-5 +<br>LiteLLM Kimi-K2-Instruct<br>for structured parsing"] & DB3[("MongoDB<br>PricingPlanSegmentContribution")]
    D5 --> D6["Experimental Pricing Recommendation Agent<br>deepresearch/experimental_pricing_recommendation.py"] & AI4["OpenAI GPT-5<br>with rabbithole thinking"]
    D6 --> AI5["OpenAI GPT-5 +<br>LiteLLM Kimi-K2-Instruct<br>for structured parsing"] & DB4[("MongoDB<br>RecommendedPricingModel")]
    E --> E1["connectors.py<br>delete_one/delete_many"]
    E1 --> E2[("MongoDB<br>Delete Operations")]
    F --> F1["connectors.py<br>list_one_markdown<br>list_all_markdown"]
    F1 --> F2[("MongoDB<br>Query Operations")]
    G["Data Models<br>datastore/models.py"] --> G1["Product"] & G2["ProductPricingModel"] & G3["CustomerSegment"] & G4["PricingPlanSegmentContribution"] & G5["CustomerUsageAnalysis"] & G6["RecommendedPricingModel"] & G7["TimeseriesData"]
    H["Utils<br>utils/openai_client.py"] --> H1["OpenAI Client"] & H2["LiteLLM Client"]
    I["Prompts<br>deepresearch/prompts.py"] --> I1["System Prompts<br>for each agent"]
    F2 --> n1["Untitled Node"]
    style A fill:#e1f5fe
    style D1 fill:#f3e5f5
    style AI1 fill:#fff3e0
    style DB1 fill:#e8f5e8
    style AI2 fill:#fff3e0
    style DB2 fill:#e8f5e8
    style AI3 fill:#fff3e0
    style DB3 fill:#e8f5e8
    style AI4 fill:#fff3e0
    style AI5 fill:#fff3e0
    style DB4 fill:#e8f5e8

Loading

Key Architecture Components

Data Layer: MongoDB serves as the central data repository with structured collections for products, customer segments, pricing models, and analysis results.

AI Integration Layer: Multiple AI models are strategically used - GPT-5 for complex reasoning, o3-deep-research for comprehensive analysis.

Agent Orchestration: The orchestrator coordinates the sequential execution of specialized agents, ensuring data flows correctly between analysis phases.

CLI Interface: Simple command-line interface provides easy access to all functionality while maintaining the sophisticated AI processing underneath.

Data Format Requirements

Before using the pricing consultant, prepare your product, pricing, and customer data in JSON format.

JSON Format

The JSON format organizes data into three main sections: product, pricing_model, and customer_segments.

Required Structure:

{
  "product": {
    "name": "Your Product Name",
    "icp_description": "Ideal Customer Profile description",
    "unit_level_cogs": "Cost of goods sold details",
    "features_description_summary": "Summary of key features and capabilities",
    "product_documentations": ["https://docs.example.com", "https://api.example.com"]
  },
  "pricing_model": {
    "unit_price": 99.0,
    "min_unit_count": 1,
    "unit_calculation_logic": "per_seat",
    "min_unit_utilization_period": "monthly"
  },
  "customer_segments": [
    {
      "customer_segment_uid": "enterprise",
      "customer_segment_name": "Enterprise Customers",
      "customer_segment_description": "Large organizations with 500+ employees",
      "number_of_active_subscriptions": 50,
      "number_of_active_subscriptions_forecast": 75,
      "usage_analyses": [
        {
          "customer_uid": "customer_001",
          "customer_task_to_agent": "Automated customer support and FAQ handling",
          "predicted_customer_satisfaction_response": 8.5,
          "predicted_customer_satisfaction_response_reasoning": "High satisfaction due to reduced response time and 24/7 availability"
        }
      ]
    }
  ]
}

Field Descriptions

Product Fields:

• name / product_name: Your product's name (required)
• icp_description: Description of your ideal customer profile
• unit_level_cogs: Detailed cost structure per unit
• features_description_summary: Key features and value propositions
• product_documentations: URLs to product docs, APIs, or resources

Pricing Model Fields:

• unit_price: Price per billing unit (default: 99.0)
• min_unit_count: Minimum units required (default: 1)
• unit_calculation_logic: How units are calculated ("per_seat", "per_usage", etc.)
• min_unit_utilization_period: Billing frequency ("monthly", "yearly", etc.)

Customer Segment Fields:

• customer_segment_uid: Unique identifier for the segment (required for CSV)
• customer_segment_name: Descriptive name for the segment
• customer_segment_description: Detailed segment characteristics
• number_of_active_subscriptions: Current subscription count
• number_of_active_subscriptions_forecast: Projected future subscriptions

Usage Analysis Fields:

• customer_uid: Individual customer identifier
• customer_task_to_agent: Description of tasks the customer performs
• predicted_customer_satisfaction_response: Satisfaction score (0-10)
• predicted_customer_satisfaction_response_reasoning: Explanation for the score

Important Notes

• Required Fields: Product name is mandatory.
• Multiple Segments: JSON supports multiple segments in an array.
• Documentation URLs: Product documentation URLs help the AI understand your product better for more accurate pricing recommendations.
• Numeric Fields: Ensure satisfaction scores, pricing, and subscription counts are valid numbers.

How to use it?

Prerequisites

# Install dependencies
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

# Set API keys
export OPENAI_API_KEY="your-openai-api-key"
export TOGETHER_API_KEY="your-together-ai-api-key"

# Optional: Set MongoDB connection
export MONGODB_URI="mongodb://localhost:27017/pricing-research"  # or your Atlas URI

MongoDB Options:

• Local: Install MongoDB and run mongod
• Cloud: Use MongoDB Atlas and set MONGODB_URI
• Docker: docker run -d -p 27017:27017 mongo:latest

Usage

# Create data from JSON
python main.py --create path/to/your/data.json

# Run pricing analysis
python main.py --orchestrator product_id

# List data
python main.py --list collection_name

# Delete data
python main.py --delete collection_name document_id