> For the complete documentation index, see [llms.txt](https://help.dscout.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.dscout.com/dscout-ai/dscout-ai-studio/ai-moderated-studies.md).
# AI moderated studies
{% hint style="info" %}
*This feature is currently in closed beta, which means it is ready for you to use, but there will be changes in functionality leading to a general release. Check back here for the latest updates, and if you have any issues or feedback, our support team is always happy to help at* [*support@dscout.com*](mailto:support@dscout.com)*.*
{% endhint %}
AI moderated studies let you conduct interview-like sessions at the speed and scale of unmoderated studies. Using your research objectives and questions as a guide, Dscout AI will dynamically ask participants questions to collect the insightful data you’d expect from a moderated session.
AI moderated studies are:
* **Continuously recorded:** Participant screens, webcams, and audio will be recorded for the entirety of the study.
* **Desktop-only:** Participants will be required to complete AI moderated studies using a desktop or laptop computer—they cannot use a mobile device.
* **Tailored to each participant:** The AI moderator adjusts its focus and depth based on participant responses (and your follow-up settings), meaning no two sessions are exactly the same but your research objectives always remain consistent.
## Best practices
You'll find thorough guidance for building an AI moderated study below, but here are some high-level best practices to keep in mind:
* **Write questions as you would in other studies:** Avoid double-barreled questions or covering more than one topic in a single question. Enable follow-ups to dig deeper into a given topic, or add another question if you have more you want to hear. For more information, see [Follow-up questions](#follow-up-questions).
* **Use Read verbatim for tighter control:** When you want the moderator to stick to a question as you've written it, enable the Read verbatim toggle. This is a **Talk out loud** question-level setting, so use it when you want questions asked word-for-word. For more information, see [Add questions](#add-questions).
* **Don't use task URLs:** Tasks as you know them in Dscout usability tests are not currently supported in AI moderated studies. This capability is coming soon.
* **Define your Research objective with analysis in mind:** Your stated research objectives will be used during study fielding (by the AI moderator) as well as during analysis (by AI summaries, themes, etc.). Keep this in mind when defining objectives initially. For more information, see [Define research objectives](#define-research-objectives).
* **Include past findings in Moderator context:** Include any past research findings in your moderator context so the moderator knows what you know, as well as what you still want to find out. For more information, see [Define moderator context](#define-moderator-context).
* Set expectations in the custom introduction: Use the custom introduction to let participants know what you'll be covering in the session, including any special instructions you want them to follow. For more information, see [Add custom introduction](#add-custom-introduction).
* **Test your study (at least) once:** To get an idea of how the AI moderator will conduct your sessions, we recommend testing your study yourself at least one time prior to launch.
## Build an AI moderated study
When building an AI moderated study, you’ll start by defining your overall research objectives. Then, you’ll add questions, combining closed-ended questions with open-ended conversation and task questions. For open-ended questions, you’ll generate question-specific moderation guides in the builder for the AI moderator to follow when conducting participant sessions.
### Add custom introduction
By default, participants will be guided through their session by the AI moderator. However, you have the option of adding a bit of custom intro text that will be presented to participants before they begin their session.
**To add custom introduction text:**
{% stepper %}
{% step %}
On the **AI moderator** tab of your study, click **Introduction**.
{% endstep %}
{% step %}
Click anywhere in the **Introduction** section. A custom message text box is displayed.
{% endstep %}
{% step %}
Enter your custom introduction text in the **Custom message** text box.
{% endstep %}
{% step %}
Click **Done**.
{% endstep %}
{% endstepper %}
Now, your custom introduction is saved and will be presented to participants at the start of their study.
### Define research objectives
On the **AI moderator** tab of your study, you’ll define your overall research objectives:
Here, clearly state what you hope to achieve in your study. Maybe you want to evaluate the intuitiveness of a new user interface. Or maybe you want to assess the perception participants have of your company’s rebrand. Whatever your goal, here are some guidelines to follow when adding them to the builder:
* Write 3–5 clear objectives. Keep each to one sentence.
* Focus on what you want participants to share, reflect on, or demonstrate.
* Keep objectives specific and distinct—avoid combining multiple goals into one.
The AI moderator will use these objectives to generate question-specific moderation guides throughout your study.
### Define moderator context
On the **AI moderator** tab of your study, you can provide context the AI moderator should keep in mind during participant sessions:
Use this field to provide any guardrails you want to the AI moderator's conversations. Maybe your company uses certain terminology, or maybe there are certain things you *don't* want the AI moderator to mention. Enter those here for the AI moderator to keep in mind when conducting sessions.
### Add questions
Once you’ve defined your research objectives, you’ll add your study questions on the **AI moderator** tab.
AI moderated studies support the following question types:
| Question | Description |
| ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Talk out loud** |
Allow participants to provide an open-ended response with the option of follow-up questions from the AI moderator.
This question type works similarly to how it does in other study types, but here you have two additional options:
Read verbatim: When enabled, the AI moderator asks the prompt exactly as written. When disabled, the AI moderator may rephrase it naturally while staying on topic. By default, this setting is Off.
Follow-up questions: Have the AI moderator ask follow-up questions in response to the participant's answer. For more information, see Follow-up questions.
|
| **Multiple choice** |
Provide participants with a question/prompt and a list of options to choose from.
Tip: Knockout logic can be applied to multiple choice questions to disqualify applicants during the screener portion of your mission. For more information, see Knockout questions.
|
| **Scale** |
Have participants rate how they feel about a statement on a predefined scale. You can customize the Start value and End value with your own labels.
Note: Scale questions do not support stim in AI moderated studies.
|
{% hint style="info" %}
**Note:** Your study’s estimated length is displayed at the top of the builder and updated as you add questions. We recommend keeping AI moderated studies to 30 minutes or less to avoid participant fatigue.
{% endhint %}
### Follow-up questions
You're in control of how dynamic or rigid you want the AI moderator to be when interacting with your participants. For example, maybe you want the AI moderator to dig a little deeper on certain questions but not on others. You can customize the number of follow-up questions the AI moderator asks on a question-by-question basis.
To customize follow-ups, use the **Follow-up questions** setting in the right sidebar when adding study questions:
When follow-ups are enabled, you have the following options:
* **Only if response is short:** Up to **one** follow-up, but no follow-ups will be asked at all if the participant's initial response is thorough.
* **1 follow-up:** One follow-up is asked regardless of thoroughness of participant's initial response.
* **2–3 follow-ups:** At least two follow-ups are asked regardless of thoroughness of participant's initial response, but no more than three follow-ups will be asked total.
You can also offer the AI moderator guidance on what kinds of follow-ups to ask. For example, maybe you want the moderator to focus specifically on asking about competitors mentioned by the participant, or maybe you want to ask for reasoning behind a participant's selection on a closed-ended question. Enter that in the **Follow-up guidance** field:
### Enable multi-language support
AI moderated studies let you scale your research in ways other study types can't support, one of which is conducting interview-style sessions in multiple languages. You can run your AI moderated study in up to 14 different languages, allowing participants to choose what works best for them. Then, responses and transcripts are automatically translated back to English for analysis.
**To enable multi-language support:**
{% stepper %}
{% step %}
In the builder, click the **Mission settings (gear)** icon. Your mission's settings are displayed.
{% endstep %}
{% step %}
Set the **Allow non-English speakers** toggle to **ON**. A language drop-down is displayed.
{% endstep %}
{% step %}
From the drop-down, select which languages you want to support. Participants will be able to choose one of these selected languages at the start of their session.
{% endstep %}
{% step %}
Click **Save**.
{% endstep %}
{% endstepper %}
Now, your language selection is saved. When you launch your study and participants join, they'll be able to select their preferred language. The entire session will be conducted in that language, including written questions and answer choices, as well as the AI moderation itself.
## Launch an AI moderated study
When you’re ready to open your study to participants, you’ll launch it from the builder.
{% hint style="warning" %}
**Warning:** You cannot edit an AI-moderated study once it’s been launched. We recommend testing your study prior to launch to see how the AI moderator interprets your moderation guides and how it reacts to a natural conversation.
{% endhint %}
Depending on your chosen recruitment approach, one of the following will happen when you launch:
* **Auto-recruit:** Your study will become available to applicants from Dscout’s Scout Panel or Partner Panels, depending on your selection.
* **Add from screeners:** Selected participants will receive an email inviting them to your study.
* **Invite via share link:** Your study’s share link will be activated and accessible to participants who have it.
* **Invite from Private Panels:** Selected participants will receive an email inviting them to your study.
## Participant experience
Regardless of recruitment approach, the participant experience in an AI moderated study will look the same.
If you've enabled multiple language options, participants will first be prompted to select their preferred language:
Then, participants will undergo a hardware setup, selecting a screen to share and granting access to their webcam and microphone:
Once their hardware is set, they’ll be welcomed to the study with any custom introduction you've added. When they're ready to start, they'll click **Begin recorded session**:
As they progress through the study, bars at the top of the page indicate the participant’s position. And at the bottom of the page, a control bar lets them mute their microphone, pause the session, flag a concerning moment, or end the session:
At the end of your study, the AI moderator will thank the participants and automatically end the session.
## Analyze an AI moderated study
To help you get to the heart of responses quickly, AI moderated studies offer two unique analysis features: summaries of open-ended questions and emerging themes for fast immersion.
### Open-ended summaries and themes
From the **Responses** tab of your study, you can view responses to closed-ended questions as you would in any other Dscout study type. However, open-ended questions are displayed slightly differently.
For each open-ended question, you’ll find the following:
* AI-generated summaries from all responses to a given question.
* AI-generated themes from all responses to a given question.
* AI-generated summaries for each individual participant (as opposed to the raw transcript). To view entire transcripts for participant responses, navigate to the full session recording view.
{% hint style="info" %}
**Note:** AI-generated summaries and themes for AI moderated studies are updated continuously as responses roll in.
{% endhint %}
**Summaries:**
**Themes:**
### Multi-lingual transcripts
If you've enabled multiple languages in your study, you can view transcripts in both English and the original language. To switch between languages, use the language toggle in the top right of the full session recording view:
### Emerging themes
AI moderated studies collect interview-depth responses at scale. But how do you *review* those responses at a speed to match? Dscout AI helps by surfacing emerging themes in automatic playlists that update with new clips as responses roll in. Themes are identified using your research objectives as inspiration as well as comments or topics that arise across participants.
The emerging themes modal opens the first time you access the **Responses** tab of your study. You can also access it again later by clicking the icon in the bottom left corner of the page:
In the modal, select a theme from the list to view a playlist of supporting clips from real participant responses:
If a playlist strikes a chord, you can share it with a colleague directly using the **Copy link** icon. Or, you can open it directly in Dscout's playlist editor by clicking the **View playlist** icon to modify it before sharing more broadly:
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://help.dscout.com/dscout-ai/dscout-ai-studio/ai-moderated-studies.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.