Release Notes: February 4, 2026

These improvements are being progressively rolled out and will be available to all customers soon.

Key Improvements

Conversation Task

Conversation Tasks have been upgraded to support more natural participant responses and higher-quality data.

Voice Input with Talk Mode

Participants can now respond using voice as an alternative to typing with Talk Mode. They can choose their preferred mode when entering a Conversation Task and switch between Talk and Type at any time.

Edit the Most Recent Response

Participants can edit their most recent message, whether typed or dictated, before continuing. When a response is edited, the AI moderator automatically updates the follow-up, keeping the conversation accurate and aligned.

Ask AI Enhancements

We have continued to enhance Ask AI to better handle time-based questions and complex filters, improving the accuracy and reliability of responses.

Time-Based Questions, Handled Automatically

Ask AI now understands date references in your questions, such as “yesterday,” “last week,” or specific dates like “December 1st.” Responses are automatically filtered to match the requested time period, making it easier to analyze recent feedback, historical trends, or changes over time without extra setup. For example:

“What budgeting patterns have emerged in the past 3 months?”

More Visible Access to Help Desk and Ask AI

The Help Desk is now easier to find in the top navigation, while Ask AI has been moved to the bottom right corner of the platform. The Ask AI panel has also been visually refined for a cleaner, more polished experience.

Additional Improvements

Unified Activity Creation Experience

Setting up activities is now more consistent and intuitive across the platform. Standard, Journal, Live Group Chat, and Live Video Interview activity creation pages now follow a consistent layout, structure, and terminology, making it easier to move between activity types with confidence.

Logic-Aware Editing Rules

We have added safeguards to protect data quality while maintaining flexibility for participants. Participants can continue to edit responses for tasks that do not affect logic. Tasks that drive logic, such as piping or visibility logic, become read-only after submission to preserve data integrity.