Skip to main content

​
Customer Management Tools

The customer management system includes tools for personalization, language preferences, memory management, and feedback collection. These tools enable persistent customer relationships and service optimization.

​
Tools Overview

​
Core Customer Tools

  • manage_customer_memory - Persistent preference storage
  • change_customer_language - UI language preferences
  • register_feedback - Structured feedback collection
  • get_customer_previous_orders - Order history and favorites

​
manage_customer_memory

​
Overview

Stores and retrieves persistent customer preferences that enhance future interactions. Memory can be scoped to a specific location or across an entire organization.

​
Operations

​
get Operation

Retrieves stored customer information: 
{
  "operation": "get",
  "memory_type": "location"
}
Response: 
{
  "success": true,
  "memory": "Vegetarian, no mushrooms, prefers thin crust pizza",
  "memory_type": "location"
}

​
set Operation

Stores new customer information: 
{
  "operation": "set", 
  "memory": "Allergic to nuts, prefers morning deliveries, celebrating anniversary this week",
  "memory_type": "location"
}
Response: 
{
  "success": true,
  "message": "Memory updated successfully for location",
  "memory": "Allergic to nuts, prefers morning deliveries, celebrating anniversary this week"
}

​
delete Operation

Removes stored memory: 
{
  "operation": "delete",
  "memory_type": "location" 
}

​
Memory Scopes

​
location (Default)

Information specific to the current restaurant:
  • Dietary restrictions for this location’s menu
  • Delivery preferences specific to this address
  • Service preferences for this location

​
businessOrganization

Information shared across all locations in a chain:
  • Universal dietary restrictions (nut allergies, vegetarian)
  • Payment preferences
  • Communication preferences

​
What to Store

​
✅ Good Memory Content

  • Dietary restrictions: “Vegetarian”, “Nut allergy”, “Gluten-free”
  • Preferences: “Likes extra spicy”, “Prefers morning deliveries”
  • Important context: “Works night shifts”, “Has mobility issues”
  • Special instructions: “Leave at door”, “Call upon arrival”
  • Personal details: “Celebrating anniversary”, “Regular customer”

​
❌ Avoid Storing

  • Order-specific information (use order draft instead)
  • Temporary information
  • Sensitive personal data beyond service needs
  • Financial information
  • Exact addresses (handled separately)

​
Security Features

​
Content Sanitization

  • Automatically limits content to 1000 characters
  • Warns about excessively long content
  • Strips unnecessary whitespace

​
Sensitive Data Detection

Automatically rejects content containing:
  • Credit card patterns: \b\d{4}[-\s]?\d{4}[-\s]?\d{4}[-\s]?\d{4}\b
  • Social security patterns: \b\d{3}-\d{2}-\d{4}\b
{
  "success": false,
  "error": "Memory content may contain sensitive data. Please remove personal identifiers."
}

​
change_customer_language

​
Overview

Updates the customer’s preferred language for UI elements, WhatsApp flows, and order summaries.

​
Supported Languages

  • fr - French
  • en - English
  • de - German
  • it - Italian
  • es - Spanish

​
Usage

{
  "language": "fr"
}
Response: 
{
  "success": true,
  "message": "Customer language updated to French",
  "language": "fr"
}

​
What Changes

Language affects:
  • WhatsApp flow interfaces
  • Button text and labels
  • Order confirmation messages
  • Payment setup instructions
  • System notifications

​
register_feedback

​
Overview

Structured feedback collection with numerical scoring for service optimization and customer satisfaction tracking.

​
Score Types

​
general_score (Required)

Overall satisfaction rating (0-100):
  • 90+ = Excellent experience
  • 70-89 = Good experience
  • 50-69 = Average experience
  • Below 50 = Poor experience

​
Optional Scores

  • food_score - Food quality rating
  • delivery_score - Delivery service rating
  • ordering_experience_score - Ordering process rating

​
Usage

{
  "feedback": "The pizza was amazing and delivered hot! Very happy with the service.",
  "general_score": 95,
  "food_score": 98,
  "delivery_score": 92,
  "ordering_experience_score": 90
}
Response: 
{
  "success": true,
  "message": "Thank you for your feedback! We really appreciate it.",
  "feedback_id": "feedback_12345",
  "scores": {
    "general_score": 95,
    "food_score": 98,
    "delivery_score": 92,
    "ordering_experience_score": 90
  }
}

​
Feedback Collection Triggers

  • After order completion
  • When customer expresses satisfaction/dissatisfaction
  • Proactive feedback requests
  • Follow-up surveys via scheduled messages

​
get_customer_previous_orders

​
Overview

Retrieves customer order history for personalization and quick reordering. This information is automatically included in conversation context.

​
Features

  • Previous orders with detailed items
  • Frequently ordered products (favorites)
  • Order statistics and patterns
  • Payment method history
  • Delivery preferences from past orders

​
Integration

Memory is automatically loaded at conversation start, so direct tool usage is typically unnecessary unless checking for updates during the conversation.

​
Integration Patterns

​
Personalization Flow

// 1. Check existing memory
{
  "operation": "get",
  "memory_type": "location"
}

// 2. Use memory to personalize recommendations
"Based on your preferences (vegetarian, thin crust), I recommend..."

// 3. Update memory with new preferences
{
  "operation": "set", 
  "memory": "Vegetarian, thin crust, discovered likes arugula pizza",
  "memory_type": "location"
}

​
Language Switching

// Customer writes in different language
{
  "language": "fr"
}

// Future interactions now in French
// WhatsApp flows appear in French
// Order confirmations in French

​
Feedback Collection

// After positive interaction
{
  "feedback": "Service was excellent, very fast delivery!",
  "general_score": 92,
  "delivery_score": 95
}

// After identifying issues  
{
  "feedback": "Food was cold when delivered, pizza quality not great",
  "general_score": 45,
  "food_score": 40,
  "delivery_score": 30
}

​
Memory Management Best Practices

​
Content Guidelines

  • Be specific: “No mushrooms on pizza” vs “Doesn’t like some toppings”
  • Include context: “Works night shifts - prefers late delivery”
  • Update regularly: Add new preferences, remove outdated ones
  • Use clear language: Avoid ambiguous descriptions

​
Scope Selection

// Location-specific dietary needs
{
  "memory": "No seafood (this location serves fish pizza)",
  "memory_type": "location"
}

// Organization-wide restrictions  
{
  "memory": "Severe nut allergy - check all ingredients",
  "memory_type": "businessOrganization" 
}

​
Privacy Considerations

  • Store only service-relevant information
  • Avoid personal details unrelated to orders
  • Regularly clean outdated preferences
  • Respect customer requests to delete information

​
Error Handling

​
Memory Operations

{
  "success": false,
  "error": "No location context available"
}

​
Language Changes

{
  "success": false,
  "error": "Language 'xx' not supported. Available: fr, en, de, it, es"
}

​
Feedback Validation

{
  "success": false,
  "error": "general_score is required and must be between 0 and 100"
}

​
Analytics and Insights

Memory and feedback data enables:
  • Personalization improvement based on stored preferences
  • Service quality monitoring through feedback scores
  • Language preference trends across customer base
  • Common dietary restrictions for menu optimization
  • Customer satisfaction patterns for service enhancement
Customer management tools create lasting relationships by remembering individual preferences and continuously improving service based on feedback, while respecting privacy and security requirements.