How to Use Custom Attributes in TBit for Automated Customer Profiling

Key Takeaways

• Custom Attributes let you define structured profile fields (text, number, boolean, date, closed list, open list) for your contacts.
• AI Inference enables your agent to automatically extract and save customer information from natural conversations -- no manual data entry required.
• Attributes power audience segmentation for targeted campaigns, letting you filter contacts by any attribute value.
• Write clear descriptions for each attribute to improve AI extraction accuracy.

What Are Custom Attributes?

Custom Attributes (called User Attributes in the TBit platform) let you define structured data fields that capture important information about your contacts. Think of them as custom profile fields: you decide what to track -- a customer's name, subscription tier, interests, birthday -- and TBit stores those values for every contact your AI agent interacts with.

The real power comes from AI Inference. When you enable it on an attribute, TBit's AI agent automatically extracts and saves that information from natural conversations. If a customer says "My name is Maria and I'm interested in the Premium plan," the agent recognizes those details and stores them without any manual data entry.

Why Custom Attributes Matter

Custom Attributes solve three problems at once:

• Automated data collection -- Your AI agent captures customer details during conversations so your team does not have to do manual data entry.
• Personalized interactions -- The agent uses stored attributes to personalize future conversations, addressing customers by name and remembering their preferences.
• Audience segmentation -- You can build dynamic audiences based on attribute values, then target those segments with campaigns.

Prerequisites

Before you begin, make sure you have:

• An active TBit account with an AI agent configured
• Admin or Editor permissions on the agent
• At least one messaging channel connected (WhatsApp, Instagram, or web chat)

Step 1: Navigate to User Attributes

1. Log in to the TBit platform at platform.tbit.app
2. Select the agent you want to configure from the sidebar
3. In the left sidebar, scroll down to the Configuration section
4. Click Settings
5. Click the User Attributes tab at the top of the page

You will see the User Attributes table. If this is your first time here, the table will be empty.

Step 2: Create Your First Attribute

Click the + Add User Attributes button in the top right corner. A dialog will appear with the following fields:

Name

Enter a clear, descriptive name for the attribute. This name will appear in the attributes table, in the contact profile, and will be provided to the AI agent as context. Good examples: "Full Name", "City", "Subscription Plan", "Birthday".

Data Type

Select the type of data this attribute will store. TBit supports six data types:

• Text -- Free-form text for names, descriptions, comments, or any string value.
• Number -- Numeric values for ages, scores, quantities, or prices.
• True/False -- Boolean values for binary states like "is a returning customer" or "has completed onboarding".
• Date -- Date values for birthdays, registration dates, or appointment times.
• Closed List -- A predefined set of options. The AI can only select from these specific values. Ideal for subscription tiers, contact types, or status fields. When you select this type, you can add options one by one and toggle whether multiple selections are allowed.
• Open List -- A flexible list where the AI can add new values based on conversations. Perfect for interests, hobbies, skills, or favorite products where you cannot predict every possible value.

AI Inference

This toggle controls whether the AI agent can automatically populate this attribute from conversations. When enabled (the default), the AI will proactively look for relevant information during conversations and save it to the contact's profile.

For example, if you have an attribute called "City" with AI Inference enabled and a customer says "I'm calling from Bogota," the AI will automatically save "Bogota" as the City value.

Disable AI Inference for attributes that should only be updated manually by your team -- for instance, an internal customer score or a manually assigned account manager.

Description

Provide a clear description that helps the AI understand what information to capture. The more specific you are, the better the AI will perform. Instead of "Customer location," write "The city or region where the customer is based, as mentioned in conversation."

Step 3: Create a Closed List Attribute

Closed List attributes are particularly useful when you need the AI to choose from a fixed set of values. Here is how to create one:

1. Click + Add User Attributes
2. Enter a name such as "Subscription Plan"
3. Select Closed List as the data type
4. In the List Options section, type an option (e.g., "Free") and click Add
5. Repeat for each option: "Basic", "Premium", "Enterprise"
6. Toggle Allow multiple selection if the contact can have more than one value
7. Add a description like "The customer's current subscription tier"
8. Click Create Variable

The attribute will appear in the table showing its options as badges.

Step 4: How AI Inference Works in Practice

Once your attributes are configured, here is what happens during a real conversation:

1. A customer messages your agent via WhatsApp, Instagram, or web chat
2. The AI agent receives the message along with the list of available profile attributes
3. As the conversation progresses, the agent identifies relevant information
4. The agent calls the update_profile_attributes tool to save the extracted values
5. The values appear in the contact's profile immediately

For example, consider these attributes: "Full Name" (Text), "City" (Text), and "Interests" (Open List). If the customer says:

"Hi, I'm Carlos from Medellin. I'm looking for information about your skincare products."

• Set Full Name to "Carlos"
• Set City to "Medellin"
• Add "skincare products" to the Interests list

No manual work required.

Step 5: View and Edit Attribute Values

You can view and edit attribute values for any contact in two places:

In the Chat Profile Panel

When you open a conversation in the Chat view, the right-side profile panel displays all attributes with their current values. You can click the edit icon to manually update any value.

In the Customer Detail View

Navigate to Clients in the left sidebar and select a contact. The Attributes tab shows all custom attributes in card format, with type indicators and AI/manual badges. You can edit values inline by clicking the pencil icon.

Step 6: Use Attributes for Audience Segmentation

Custom Attributes become even more powerful when used for audience segmentation in campaigns:

1. Navigate to Campaigns in the left sidebar
2. Create a new campaign or edit an existing one
3. When configuring the audience, click Create Dynamic Audience
4. In the filter builder, you will see a Profile Attributes section listing all your custom attributes
5. Build filter conditions like "Subscription Plan equals Premium" or "Interests contains skincare"
6. TBit will automatically calculate the matching contacts

This lets you send targeted WhatsApp template campaigns to precisely the right audience based on data your AI agent collected automatically.

Best Practices

Naming Conventions

• Use clear, human-readable names: "Full Name" instead of "full_name"
• Be consistent across attributes: if you use "Customer" in one name, use it in others
• Keep names concise but descriptive

Description Quality

• Explain exactly what information to extract
• Give examples of how the information might appear in conversation
• Specify any formatting preferences (e.g., "Store the full phone number including country code")

Data Type Selection

• Use Closed List when you have a finite, known set of values and you want to guarantee data consistency
• Use Open List when values are unpredictable -- like interests, product preferences, or skills
• Use Text for unique identifiers like names, email addresses, or free-form notes
• Use Boolean for simple yes/no flags

AI Inference Strategy

• Enable AI Inference for attributes that customers naturally mention in conversation (name, location, interests)
• Disable AI Inference for internal-only attributes (customer score, account manager, internal notes)
• Start with a few key attributes and expand gradually -- too many attributes can reduce AI accuracy

Summary

Custom Attributes in TBit transform your AI agent from a simple chatbot into an intelligent data collection system. By defining the right attributes with clear descriptions, your agent can automatically build rich customer profiles during natural conversations. Those profiles then power personalized interactions and targeted campaigns, creating a virtuous cycle of better data leading to better customer experiences.

To get started, navigate to Settings > User Attributes, create a few attributes with AI Inference enabled, and let your agent do the rest.