Customer Segment

Action ID: customer_segment

Description

Generate customer segment from Amazon Personalize.

Input Parameters

Name	Type	Required	Default	Description
prompt	string	✓	-	The prompt to use for segment generation
car_model	string	-	COROLLA CROSS PREMIUM - 2022	The car model for segmentation

View JSON Schema

{
  "description": "Customer segment node input.",
  "properties": {
    "prompt": {
      "description": "The prompt to use for the question answering.",
      "title": "Prompt",
      "type": "string"
    },
    "car_model": {
      "default": "COROLLA CROSS PREMIUM - 2022",
      "description": "The car model to use for the question answering.",
      "enum": [
        "YARIS CROSS HYBRID - 2021",
        "COROLLA CROSS PREMIUM - 2022",
        "RAV4 HYBRID - 2022",
        "HARRIER HYBRID PREMIUM - 2022",
        "COROLLA ALTIS STANDARD - 2023",
        "COROLLA ALTIS ELEGANCE - 2023",
        "CAMRY HYBRID STANDARD - 2024",
        "CAMRY HYBRID ELEGANCE - 2024",
        "SIENTA HYBRID ELEGANCE - 2021"
      ],
      "title": "Car model",
      "type": "string"
    }
  },
  "required": [
    "prompt"
  ],
  "title": "CustomerSegmentInput",
  "type": "object"
}

Output Parameters

Name	Type	Description
data	array	Array of customer segment data

View JSON Schema

{
  "description": "Customer segment node output.",
  "properties": {
    "data": {
      "items": {},
      "title": "Data",
      "type": "array"
    }
  },
  "required": [
    "data"
  ],
  "title": "CustomerSegmentNodeOutput",
  "type": "object"
}

How It Works

This node uses Amazon Personalize to generate customer segments based on a natural language prompt and optionally a car model. The system analyzes customer behavior patterns, preferences, and demographics to identify groups of customers who match the specified criteria. The prompt guides the segmentation logic, and the car model filters results to customers interested in specific vehicle types.

Usage Examples

Example 1: Find High-Value Customers

Input:

prompt: "Find customers who have made multiple purchases in the last 6 months"
car_model: "CAMRY HYBRID ELEGANCE - 2024"

Output:

data: [
  {"customer_id": "C12345", "purchase_count": 3, "total_value": 85000},
  {"customer_id": "C67890", "purchase_count": 2, "total_value": 120000}
]

Example 2: Target First-Time Buyers

Input:

prompt: "Identify potential first-time car buyers"
car_model: "YARIS CROSS HYBRID - 2021"

Output:

data: [
  {"customer_id": "N00123", "age_group": "25-30", "interest_score": 0.85},
  {"customer_id": "N00456", "age_group": "22-28", "interest_score": 0.78}
]

Example 3: Segment by Engagement

Input:

prompt: "Find customers who engaged with our emails but haven't purchased"
car_model: "COROLLA CROSS PREMIUM - 2022"

Output:

data: [
  {"customer_id": "E99001", "email_opens": 15, "last_engagement": "2024-11-20"},
  {"customer_id": "E99002", "email_opens": 8, "last_engagement": "2024-11-22"}
]

Common Use Cases

• Targeted Marketing: Identify customer segments for personalized marketing campaigns

• Product Recommendations: Find customers likely to be interested in specific car models

• Churn Prevention: Identify at-risk customers who may need re-engagement

• Upsell Opportunities: Find customers ready for premium model upgrades

• Campaign Optimization: Segment audiences for A/B testing and campaign targeting

• Customer Insights: Analyze customer behavior patterns across different demographics

• Lead Scoring: Prioritize leads based on likelihood to purchase specific models

Error Handling

Error Type	Cause	Solution
Invalid Prompt	Prompt is too vague or malformed	Provide specific, clear segmentation criteria in the prompt
No Segments Found	No customers match the criteria	Broaden your search criteria or try different car models
Model Not Found	Car model value doesn't match enum options	Use one of the predefined car model options from the list
Personalize Error	Amazon Personalize service error	Check AWS Personalize status and configuration
Empty Data	No customer data available for segmentation	Ensure customer data has been imported into Personalize
Timeout Error	Query took too long to process	Simplify the prompt or reduce the scope of segmentation

Notes

• Car Models: The node supports specific Toyota car models. Select the model most relevant to your campaign.

• Prompt Engineering: Clear, specific prompts yield better segmentation results. Include timeframes, behaviors, or demographics.

• Amazon Personalize: This node requires a configured Amazon Personalize campaign with customer data.

• Data Privacy: Ensure customer data handling complies with privacy regulations (GDPR, CCPA, etc.).

• Segment Size: Very specific prompts may return small segments. Balance specificity with reach.

• Real-Time Updates: Segments reflect the current state of data in Amazon Personalize.

• Cost Considerations: Amazon Personalize charges based on data processing and recommendations generated.