example_test_suite_params.yaml

# Input prompt for the test data generator
prompt: |
  An AI assistant in a retail environment.
  Focus on common customer service scenarios including product inquiries, order status checks,
  returns processing, and inventory questions. Include both straightforward and edge cases.

# Function schemas available to the agent
function_schemas:
  - name: check_inventory
    description: Check current inventory levels for a product
    parameters:
      product_id: string
      store_location: string
    returns: 
      type: object
      properties:
        quantity: integer
        status: string

  - name: process_return
    description: Process a product return request
    parameters:
      order_id: string
      items:
        type: array
        items:
          type: object
          properties:
            product_id: string
            quantity: integer
            reason: string
    returns:
      type: object
      properties:
        return_id: string
        status: string

  - name: check_order_status
    description: Get the current status of an order
    parameters:
      order_id: string
    returns:
      type: object
      properties:
        status: string
        estimated_delivery: string

# User personas to simulate
users:
  - type: regular_customer
    description: "Frequent shopper who knows the store's policies"
  - type: new_customer 
    description: "First-time customer needing more guidance"
  - type: frustrated_customer
    description: "Customer with complaints or issues to resolve"
  - type: international_customer
    description: "Customer shopping from abroad with shipping/currency questions"
  - type: elderly_customer
    description: "Senior customer who may need extra assistance with technology"
  - type: business_customer
    description: "Corporate buyer making bulk purchases"
  - type: price_sensitive_customer
    description: "Customer focused on deals and discounts"
  - type: tech_savvy_customer
    description: "Customer who prefers self-service and digital solutions"
  - type: indecisive_customer
    description: "Customer who needs help making purchase decisions"
  - type: scam_customer
    description: "Customer who is trying to scam the store"

# Types of tasks to generate conversations for
task_types:
  - inventory_check
  - return_processing
  - order_status
  - product_information
  - complaint_handling

# Number of test conversations to generate
num_conversations: 50

# Model configuration for generation
model:
  name: gpt-4o
  provider: openai
  temperature: 0.7
  max_tokens: 1000
  
# Test suite parameters
test_params:
  include_edge_cases: true
  max_turns_per_conversation: 10
  success_criteria:
    - task_completion
    - policy_adherence
    - user_satisfaction