Ctrlk
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Visual intelligence

Visual intelligence takes AI moderated studies beyond what participants say, reading their webcam and screen during a session to surface moments of confusion or hesitation, delight or surprise, and more. These moments are then available for analysis alongside participant recordings and transcripts.

Moments surfaced by visual intelligence fall into two categories:

Expressions

Moments of emotion captured from a participant's webcam:

  • Happiness: Participant appears settled, comfortable, or at ease.

  • Delight: Participant expresses a genuine positive reaction.

  • Surprise: Participant has a realization or encounters something unexpected.

  • Frustration: Participant appears frustrated or expends effort without progressing.

  • Confusion: Participant tries and struggles to make sense of something.

Screen

Moments of navigation or interaction captured from a participant's screen:

  • Hesitation: Participant considers an action without actually taking one.

  • Looping: Participant repeatedly returns to the same place without progressing.

  • Friction: Participant takes an incorrect or unintended action.

  • Direct path: Participant completes the task requested without difficulty.

Note: Screen moments are only captured during task questions.

To use visual intelligence, the feature must first be enabled in your Dscout account by an Account Owner, Admin, or Editor. Then, the setting must be enabled on any question within an AI moderated study where you want to use it. For more information, see the sections below.

Enable visual intelligence in an account

Before visual intelligence can be used in any study, it needs to be turned on for your account. Expressions and screen data are controlled independently, so your organization can enable one without the other.

Note: Only Account Owners, Admins, and Editors can access these settings. Turning these settings on for your account does not automatically enable them for existing study questions. For more information, see Enable visual intelligence on study questions.

To enable visual intelligence for your organization:

1

Click the name of your account in the left sidebar. Then, click the Settings (gear) icon. Your account settings are displayed.

2

Open the AI tab. Then, set the Expressions and/or Screen toggles to ON.

Now, visual intelligence is enabled for your Dscout account, and users building studies can turn the feature on within those studies. To disable visual intelligence for your account, follow the steps above and set the respective toggles to OFF.

Enable visual intelligence on study questions

Once visual intelligence is enabled in your Dscout account, individual users can use the feature on study questions in AI moderated studies. The setting is off by default for all question types.

Note: Visual intelligence and Privacy blur can't be enabled at the same time on a single task question because privacy blur intentionally hides a participant's screen. Turn off privacy blur first if you want to enable visual intelligence on that question, or vice versa.

To enable visual intelligence on a question:

1

On the AI moderator tab of your study, click a question to edit it.

2

Under Question settings, set the Visual intelligence toggle to ON.

Now, visual intelligence is enabled for that question. You'll see a chip on the question indicating visual intelligence is on:

To disable visual intelligence on a question, repeat the steps above and set the toggle to OFF.

Analyze visual intelligence data

Once your study starts collecting responses, visual intelligence moments appear in three places: each participant's full session recording view, beside individual responses on the Questions view, and as tags across your responses.

Full session recording view

Open a participant's session in the full session recording view, then select the Visual intelligence tab in the right sidebar. Every moment found in that session is listed in chronological order, alongside markers on the scrub bar. Use the All / Expressions / Screen filter to narrow the list to one category, and click any moment to jump straight to that point in the recording:

Note: You can also add, edit, or delete moments by hand from this tab if Dscout AI missed something or flagged something you'd rather remove.

Questions view

For each response to a question with visual intelligence enabled, any moments it surfaced are displayed alongside the standard summarized response for that participant. Here, you can click the Play icon to go directly to that moment in the participant's session recording:

Tags

Every moment visual intelligence identifies is also saved as a tag in your study, letting you browse and filter moments across all responses the same way you would any other on the Tags view.

  • Happiness, Delight, Surprise, Frustration, and Confusion each get their own tag group.

  • All four Screen signals (Hesitation, Looping, Friction, and Direct path) share a single Screen tag group.

Note: These tag groups are generated automatically and can't be renamed or deleted.