Skip to main content
Your AI agent collects customer details, such as skin type or shoe size, from conversations. Describe your fields, then review uncertain values in the customer’s Additional data card.
High-confidence values are saved automatically and can replace an existing value. Only uncertain values need approval.
1

Describe your data fields

The agent uses field names and optional descriptions to identify what to collect.
  • Open Customers and open a customer’s profile.
  • In the Additional data card, add a new field or edit an existing one.
  • Fill in Description (optional) (up to 255 characters). For example: “Customer’s skin type or shoe size”. Additional data card with the Add new field and Edit value actions Data field editor with field name, data type, and optional description
Data fields are managed from the customer profile. The Collect data skill in AI Studio links to them with Manage.
2

Review suggestions

Open Additional data from the customer’s profile in Customers, or from Customer Details in bitChat live chat.Each suggestion shows as a purple pill, for example Suggested: add Oily. If the field already has a value, the current value appears above the pill.
  • Hover over the pill to see the customer’s message behind the suggestion, for example: “my skin gets oily by noon”.
  • Click the check icon to approve the suggestion, or the X icon to reject it.

Values saved by the AI agent

A value the AI agent saved automatically, or one you approved, shows an AI badge. The badge means your AI agent collected it from a customer conversation. If you edit the value, your edit replaces it.
A returning customer’s Preferred size field contains S. During a new conversation, they say, “I now wear size L and prefer a looser fit.” Keeping those details on their profile helps the next agent avoid asking the same questions again.Create separate fields for the preferences you need, such as Preferred size and Preferred fit, and describe what each field should capture. These are example fields you define for your business, not default fields.
  • If the AI has high confidence in the new value, it updates the field automatically, including replacing an existing value.
  • If confidence is low, it suggests a value for your team to approve or reject. A change to an existing preference does not always require approval.
  • After the conversation, check the customer’s Additional data card and review any suggestions against the customer’s message.
Use customer data fields for individual details such as a preferred size. To group customers by product interests for a campaign, see AI Tagging.