QuestionPunkQuestionPunk
Start free

Run a usability test with screen sharing

Set up an AI-moderated usability test where respondents share their screen, complete tasks, and think aloud.

A Usability test question runs an AI-moderated usability session. Respondents share their screen, complete guided tasks on a website or app, and think aloud while the AI moderator observes, asks follow-up questions, and collects post-task ease ratings. Results include task outcomes, navigation analytics, and per-task video clips when screen recording is enabled. A usability test can also run on a Figma prototype: paste the prototype link, pick the start and goal screens, and respondents use the prototype embedded in the study, with its screen changes noted alongside the recording.

Last reviewed
aiusabilityscreen-shareresearchux15-30 minutes to set upIntermediate to AdvancedUX researchersProduct teamsDesigners

Before you start

Prepare the task scenarios, target URLs, and any micro-steps you want to track. If you are testing a native app, have the Android package name, iOS URL scheme, and iOS App Store ID ready.

Steps

  1. Add a Usability test question

    In the survey editor, click Add Question. Then choose Usability test from the AI category. This question type is built for AI-moderated usability sessions with screen sharing and task tracking.

  2. Add tasks

    A Usability test always uses guided tasks with screen sharing, so there is nothing to turn on. Open the question. Then add the tasks you want respondents to complete.

  3. Set up task scenarios

    Add up to 20 task scenarios. For each task, add the scenario text that the AI reads aloud, a target URL for a website, app, or prototype, optional micro-steps to track sub-task completion (up to 50 per task), a time budget (30 seconds to 45 minutes, default 5 minutes), and a stuck threshold (10 seconds to 60 minutes with no progress before the AI intervenes, default 120 seconds). Mark warmup tasks with the warmup flag so they are excluded from analysis.

  4. Choose task ordering

    Enable randomize task order to vary task sequence across respondents and reduce order effects. You can also randomize step order within tasks.

  5. Set think-aloud and intervention behavior

    Think-aloud reprompting is on by default. If the respondent is silent for 20 seconds, the AI gives a neutral "keep talking" prompt. You can change this to 5-120s. Set Moderator style to Coach (AI reads each task, guides step by step, answers questions, and moves on when done) or Silent observer (AI reads each task and watches quietly, speaking only to advance or debrief). Under When the participant asks for help, choose Deflect, Minimal answer, or Don't respond. The reassurance intro is on by default. You can also add up to 10 post-task questions per task and up to 5 wrap-up questions after all tasks.

  6. Use screen sharing

    Screen sharing is always on for a Usability test question. On desktop, respondents share their screen through the browser's screen capture. On Android, the mobile app uses native screen sharing.

  7. Check the results

    The results show a per-task breakdown: outcome distribution (success/partial/fail/abandoned), average duration, navigation friction (revisits, oscillations, back-navigations), and per-step completion rates with stuck and wrong-turn indicators. When screen recording is enabled, per-task video clips are cut from that recording.

Details and limitations

Usability testing is a dedicated question type in QuestionPunk that combines AI-moderated voice interviews with screen recording and guided task tracking. Respondents speak their thoughts while completing tasks. The AI moderator adapts its questions based on what it observes. Screen sharing works in desktop browsers and in the Android app.

Each task scenario defines what the respondent should accomplish and on which URL. You can also set time limits and stuck detection. The AI looks for signs of difficulty and can coach or stay silent, depending on your moderator style.

Think-aloud reprompting helps respondents keep narrating their thought process. If they go silent, the AI gives a neutral prompt after a configurable silence threshold (default 20 seconds).

A Single Ease Question (SEQ, 1-7 scale) is asked after each task by default. You can turn it off. This gives you a standardized ease metric alongside qualitative observations.

Use the outcome, navigation, and per-step results to find where people got stuck, took wrong turns, or gave up.

Navigation analytics track every screen visit. They compute revisit counts, oscillation patterns (A→B→A), and back-navigation frequency per task. This helps you spot navigation friction without reading logs.

When screen recording is enabled, QuestionPunk can cut per-task video clips from the full recording. Check recording settings before the session if you need clips to share with stakeholders.

Counterbalanced task ordering randomizes the sequence across respondents. This helps control fatigue and learning effects in within-subjects designs.

Native app testing: Task targets can be set to website, app, or prototype. For native app targets, configure an Android package name, iOS URL scheme, and iOS App Store ID to launch the respondent directly into the app being tested.

Per-language task text overlays let you translate scenario descriptions, step instructions, and post-task questions for multilingual usability studies.

Match recruitment to the supported device. Desktop browsers can share a screen; Android browser participants need the supported app handoff. An iPhone or iPad browser cannot provide the required screen share. A field for an iOS app URL does not establish iOS browser screen-sharing support. Pilot the complete task on the intended device before inviting participants.

Screen sharing and recording are different. Enable recording before the session if you need a saved recording or task clips. Explain what will be captured, and ask participants to close unrelated windows and notifications before sharing. For permission or interrupted-audio problems, use the audio and screen-sharing guide.

Check the result

Check the task scenarios, timing, and moderator behavior in Preview before recruiting. After completed sessions, review task outcomes and navigation metrics. Video clips require screen recording to be enabled.

If something goes wrong

If the session does not match your study plan, open the question and check the task scenario details, including the target URL, time budget, and moderator style.