Introduction to Segmentation in Zowie

What is Customer Segmentation in Zowie?

Customer segmentation in Zowie AI Agent is a powerful feature that allows you to categorize your customers into distinct groups based on specific criteria. These segments serve as organizational units that enable targeted content delivery, personalized responses, and customized automation workflows.

Segmentation works as a fundamental layer within Zowie's modular architecture, directly affecting how the Decision Engine and Reasoning Engine interact with customers. By implementing segmentation, you can ensure that your AI Agent delivers the most relevant information to different customer groups, improving both response accuracy and customer satisfaction.

Benefits of Implementing Segmentation

Implementing segmentation in your Zowie AI Agent provides several key advantages:

1. Personalized Customer ExperiencesDeliver tailored responses that address the specific needs and concerns of different customer groups.

   Example: A retail company segments customers into "Premium" and "Standard" tiers. When a Premium customer asks about shipping options, the AI Agent automatically references policies with expedited shipping options and complimentary services, while Standard tier customers receive information about regular shipping options with available upgrades. This personalization happens seamlessly without customers needing to specify their tier in each interaction.

2. Optimized Knowledge ManagementAssign policies to relevant segments, ensuring that the AI Agent references only the most pertinent information when responding to inquiries.

3. Improved Response AccuracyReduce confusion by creating clear boundaries between different service offerings, pricing tiers and other policies.

4. Enhanced AnalyticsGain deeper insights into how different customer groups interact with your AI Agent and which segments require additional knowledge resources.

How Segmentation Impacts Conversation Routing and Personalization

Segmentation directly influences several key components of the Zowie platform:

• Knowledge Retrieval:The Reasoning Engine uses segment data as a filtering mechanism when retrieving relevant policies from the Knowledge base. This ensures that customers receive information specifically applicable to their segment.

• Multi-Channel Distribution:Messages across different communication channels can be routed and prioritized based on segment attributes, ensuring consistent experiences regardless of how customers contact your support.

Step-by-Step Segmentation Configuration

Accessing the Segmentation Interface

To start configuring segmentation in Zowie navigate to Workspace > Segmentation in the main navigation menu

The Segmentation interface provides a comprehensive view of both Properties (user characteristics) and Segments (user groups defined by properties).

Note: Segmentation operates on the workspace level, meaning that user segments and properties set here will be accessible to configuration across all AI agents in your workspace. This unified approach ensures consistent customer experiences regardless of which AI agent is handling the interaction.

Understanding Properties and Segments

When you first access the Segmentation section, you'll see two primary lists:

1. Properties - Characteristics that describe your users
2. User Segments - Groups of users defined by specific property values

Default Properties

Zowie provides several pre-configured properties to help you get started:

• Authenticated: Used with JWT authentication (contact your Customer Success Manager for setup)
• Channel: Identifies which platform users are coming from (Messenger, website, mobile app, etc.)
• Current Website URL: Captures the page where the interaction occurs
• Email: User's email address
• Locale: User's current country based on browser information
• Name: User's name
• Phone number: User's contact number
• Time zone: User's local time zone

Note: These predefined properties cannot be removed from the system.

Default Segments

Zowie also creates default segments for each supported channel (e.g., a segment for all users writing from Messenger).

Note: These predefined segments cannot be removed from the system.

Creating Custom Properties

To create a new user property:

1. Click "Create new" in the top right corner of the Properties section
2. Fill out the details:
   • Name: Display name for the property (e.g., "Pet Preference")
   • Key: Technical identifier for the property
   • Type: Select the appropriate data type (String, Number, Boolean, Timestamp)
   • Allow editing this property in Inbox: Toggle on if you want human agents to be able to modify this value during conversations
     Attention: Ensure you use the correct text format in the Key field.

Gathering Property Data

You can collect property values from several sources:

• Survey: Configure customer surveys directly in the Zowie platform
• Widget code: Add properties directly to your chat widget code (may require technical assistance)
• Zowie Chat: Automatically collects basic information like IP address, browser locale, and current website URL
• Authentication: For the Authenticated property, contact your Customer Success Manager about JWT authentication

Example: Creating a "Sports Preference" Property

Let's say you want to segment customers based on their preferred sport activity:

1. Create a property named "Sports Preference" with an appropriate key
2. Set the type to "String"
3. To gather this data, create a survey that asks users about their favorite sport
4. In the survey configuration, enable "Store in Zowie as user's property" in the Integrations tab
5. Create an automation to present this survey to users

Important: Always ensure the property name in your survey configuration matches the one in your property configuration (excluding any special characters).

Creating Custom Segments

Once you've defined your properties, you can create segments based on them:

1. Click "Create new" in the top right corner of the Segments section
2. Fill out the form:
   • Name: Display name for the segment (e.g., "Cat Lovers")
   • Description: Brief explanation of the segment's purpose
   • Property: Select the property to use for segment criteria
   • Condition: Define the rules for segment membership (equals, contains, greater than, etc.)
   • Value: Specify the property value that qualifies users for this segment

You can create multiple segments based on the same property or combine multiple properties to define more specific user groups.

Example: Creating "Running Enthusiasts" and "Cycling Enthusiasts" Segments

Building on our sports preference example:

1. Create a segment named "Running Enthusiasts"
   • Select the "Sports Preference" property
   • Set condition to "equals"
   • Set value to "Running"
2. Create another segment named "Cycling Enthusiasts"
   • Select the "Sports Preference" property
   • Set condition to "equals"
   • Set value to "Cycling"

These segments will automatically include users based on their survey responses, allowing you to deliver personalized product recommendations and content to each group.

Creating Segment-Specific Knowledge

A key advantage of segmentation in Zowie is the ability to assign specific knowledge policies to relevant segments, ensuring customers receive information tailored to their characteristics. Here's how to create segment-specific knowledge:

1. Navigate to the Knowledge section in your AI Agent navigation
2. Create or select an existing Policy that should be segment-specific
3. Click "Edit" to modify the policy settings
4. Locate the "Segments" field in the policy configuration panel
5. Assign the policy to one or more segments by selecting them from the dropdown
6. Save your changes to activate segment-based policy filtering

When the AI Agent responds to a user query, it will now consider the user's segment when retrieving relevant policies. This ensures that customers only receive information applicable to their specific situation.

Example: Sports Equipment Knowledge Segmentation

A sports retailer might create separate warranty policies for different customer segments:

• "Premium Warranty Policy" assigned to the "Elite Athletes" segment
• "Standard Warranty Policy" assigned to the "Casual Athletes" segment

When a customer asks about warranty coverage, the AI Agent automatically provides the appropriate policy based on the customer's segment.

Best Practices for Effective Segmentation

Guidelines for Creating Meaningful Customer Segments

1. Start with Business Objectives: Define segments based on your specific business goals rather than arbitrary categories. Each segment should serve a clear purpose in your customer support strategy.
2. Use Data-Driven Approaches: Base your segmentation on actual customer data and interaction patterns rather than assumptions about your customer base.
3. Maintain Clear Segment Names: Use consistent, descriptive naming conventions that make segment purposes immediately clear to all team members.
4. Document Segment Definitions: Create comprehensive descriptions for each segment that outline:
   • Target customer characteristics
   • Typical support needs
   • Relevant policies and processes
   • Special considerations for this segment
5. Establish a Segment Hierarchy: Organize segments in logical groups to maintain structure as your segmentation strategy grows more complex.

Strategies for Balancing Segment Specificity and Coverage

1. Avoid Over-Segmentation: Creating too many narrow segments can lead to:
   • Excessive maintenance burden
   • Duplicate content across segments
   • Increased risk of knowledge gaps
   • Difficulty managing segment interactions
2. Use Progressive Segmentation: Start with broader segments and gradually refine them as you gather more customer data and insights.
3. Implement Segment Overlap Rules: Decide how to handle customers who might qualify for multiple segments:
   • Primary/secondary segment designation
   • Segment prioritization hierarchy
   • Composite response generation using multiple segment policies

Avoiding Common Segmentation Pitfalls

1. Inconsistent Policy Assignment: Ensure all relevant policies are appropriately tagged with segment information to prevent knowledge gaps for specific customer groups.
2. Static Segmentation: Avoid creating rigid segments that don't adapt to changing customer characteristics or business needs.
3. Excessive Rule Complexity: Complicated segment assignment rules can become difficult to maintain and may produce unexpected results. Aim for clarity and simplicity.
4. Neglecting Segment Testing: Always validate that segments work as expected before deploying them in live customer interactions.
5. Ignoring Segment Evolution: Customer groups change over time. Regularly review segment definitions to ensure they remain relevant.