Moderator settings
The AI Moderator’s persona comes from two places: the tone you write in the research framing (see Define the study), and the name, avatar, language, and voice you configure here.
Companion settings are in the Conversation Companion card in Panel 3 (Preview & Test). Language settings are in the collapsible Language Settings section at the bottom of the Frame the Study card in Panel 1.
Companion name
Section titled “Companion name”Enter a name for your Conversation Companion (required, maximum 50 characters). It defaults to Eve.
The companion’s name is shown to participants during the conversation. You can also reference it in the conversation introduction and other supported text fields using the {{companionName}} placeholder.
Companion avatar
Section titled “Companion avatar”Upload an avatar image for your Companion. Accepted formats: JPG, JPEG, PNG, GIF, BMP. Maximum file size: 5 MB.
Drag the file onto the upload area or click to browse. The current avatar is shown as a preview above the name field.
The RESET button restores the companion name to Eve and the avatar to the default EVE image. It doesn’t touch any other conversation settings. Save after resetting to preserve the change.
Conversation language
Section titled “Conversation language”Select language the AI Moderator should use during the conversation — for example, English (Australia). Only one language can be selected at a time.
The language setting changes how the AI Moderator writes its questions and responses. Set this before writing your probe and prompt text — and write your briefing fields in the same language you want the conversation to run in.
This setting also interacts with the survey language in Survey Shaper to determine whether automatic translation is applied when the conversation is distributed. See Language and survey translation.
Survey Shaper integration
Section titled “Survey Shaper integration”The Survey Shaper Integration section is collapsible — click the label to expand it. It has two settings:
Voice activation
Section titled “Voice activation”Toggle Enable AI voice prompts and voice response recording to activate voice for the conversation.
When voice is enabled, the following options appear:
- Gender — multi-select filter to narrow down available voices by gender.
- Age Group — multi-select filter to narrow down available voices by age group.
- Select Voice — choose a specific voice from the filtered list of available Azure voices. The list loads from the API — you’ll see a loading indicator while it fetches.
- Preview Voice — plays a sample of the selected voice. The button shows Playing… while the sample plays.
Frequently asked questions
Section titled “Frequently asked questions”Does every conversation need a companion name?
A name is required — but the field defaults to Eve, so you don’t need to change it if you’re happy with the default. The name is shown to participants during the conversation, so choose something that fits your research context.
Can I change the companion name after publishing?
Yes. Edit the companion name, save, and publish again. Participants in future sessions will see the new name. Completed sessions won’t be affected.
Does the companion name appear in the conversation transcript or data?
The name is the persona shown to participants during the live conversation. Check your analytics and reporting setup to see how responses are labelled in any downstream data.
What does RESET do exactly?
It restores the companion name to Eve and replaces the avatar with the default EVE image. It doesn’t touch any other conversation settings. You’ll need to save after resetting to preserve the change.
Can I run the same conversation in multiple languages?
Only one language can be selected at a time per conversation. If you need to run the same study in multiple languages, create a separate conversation for each language and configure the appropriate language settings in each.
What’s the voice list based on?
Available voices are pulled from the Azure Cognitive Services voice library. The list is filtered by your gender and age group selections — clear those filters to see all available voices. Some voices may take a moment to load.
Can I preview a voice before selecting it?
Yes. Select a voice, then click Preview Voice to hear a sample. To hear how it sounds in context, run a preview session in Panel 3 as well.
Does voice activation work in the conversation preview?
The preview in Panel 3 runs against the current draft configuration, including voice settings. Enable voice and run a preview session to test the experience end to end before publishing.
Common issues and solutions
Section titled “Common issues and solutions””My avatar image won’t upload”
Section titled “”My avatar image won’t upload””Check the file format — the companion avatar accepts JPG, JPEG, PNG, GIF, and BMP only (WEBP is not supported here). Also check the file size is under 5 MB.
”RESET changed both the name and avatar when I only wanted to change one”
Section titled “”RESET changed both the name and avatar when I only wanted to change one””RESET restores both the name and the avatar together. If you want to keep a custom avatar but change the name (or vice versa), edit the fields directly rather than using RESET.
”The voice selection list is empty or not loading”
Section titled “”The voice selection list is empty or not loading””The voice list is fetched live from the API — you’ll see a loading indicator while it loads. If the list stays empty, check your network connection and try refreshing. The gender and age group filters narrow the list — clear all filters to see all available voices.
”I enabled voice but participants don’t hear anything”
Section titled “”I enabled voice but participants don’t hear anything””Voice activation requires the Survey Shaper integration to be configured for your account. If voice is toggled on but not working, check that the integration is set up correctly and that Enable AI voice prompts and voice response recording is saved as part of the published conversation.