See how it works: Prepare text, voice or screen interviews

Interview by screens

The AI guides screen by screen and asks for clicks.

2 minutes read

On this page

Define a script, response mode and visual context.

Purpose

Collect explanations in a conversation with clear limits.

When to use it

To understand reasons through connected questions without a human moderator.

Before you start

  • Reviewed script and language
  • A microphone for voice
  • Imported screens for visual mode

Example

Explain what you expect to find on this screen and indicate where you would click.

Step by step

  1. Write the script. Add AI interview. Set context, topics, maximum duration and follow-ups to keep the conversation focused. One question per topic · Follow-up limit.
  1. Choose response mode. Select text or voice for your audience. For voice, choose turns or real time if available and test the intended browser. Turns · Real time when available.
  1. Configure screens if needed. In Screens mode, select the import, scenario and question for each screen. Choose ordered script or navigable links; enable click requests when part of the task. Script · Links · Click request.
  1. Rehearse the conversation. Preview simulates the conversation and disables live voice. Publish and open the participant link to test voice, silence and permissions with headphones. This can generate a real response; separate it when interpreting collection. Intended text and voice modes · An available exit.
  1. Read the turns. In Results, open the interview transcript and listen to available audio. For screens, relate comments to what was visible and the click made. Speech · Screen · Action.

Validate before collecting responses

The side preview lets you inspect a block; for Figma it is static. To validate draft rules and order, use Test full flow, without creating sessions or taking quota slots. Figma, recording, voice and AI follow-up require the published link. Test published opens the real /r/[token] link: it can create responses and consume quota. Separate these tests when analyzing your collection.

How to read the result

  • Turn transcripts and available audio
  • Screen context and routes when configured

Cautions and limitations

  • Real-time voice depends on connectivity and availability; rehearse fallback behavior.
  • Microphone capture does not mean the screen was also recorded.

Continue with the interactive guide

Prepare text, voice or screen interviews

Was this helpful?

Is something missing? Write to contato@uxtap.app.