Skip to main content

Voice

Experience the future of customer research with HEARD Voice.

Updated over a week ago

Our goal with "voice" was to create a seamless experience that feels natural and lifelike to research participants while also being capable of managing many different scenarios within the interview. After extensive testing, we have launched a Beta version. This version requires users to click to talk rather than dynamically responding—but it works reliably, ensures data integrity, and avoids frustrating customer experiences.

In this study, we will go over what Voice interviews are, how to set them up, what the participant experiences and how to understand your insights.

How Voice Interviews Work

This Voice feature will allow you to conduct AI-moderated interviews using natural speech, creating more engaging and conversational experiences with your research participants.

Your customers speak directly to an AI moderator, which asks your discussion questions and intelligent follow-ups based on their responses. The AI captures everything in a transcript while generating the same insights and analysis you expect from HEARD.

Setting Up Voice Interviews

Choosing "Response Mode"

  1. Setup up your study in the same way you normally would.

  2. On the "Settings" tab, you'll see an option for Response Mode. These are the three options

    1. Text - Traditional text-based interviews (this is the default experience)

    2. Voice - Participants interact with your study through speech

    3. Voice or Text - Participants choose their preferred method when they start.

Once your study has responses, you can change the response mode.

Selecting Your AI Voice

If you choose a voice-enabled modality:

  1. Click on the voice selection dropdown

  2. Preview different AI voice options

  3. Select the voice that best fits your study's tone and audience

  4. The selected voice will be used consistently for all interviews

Moderator Messages

  1. After selecting the AI Voice fill out the "Moderator Messages and Image" section of HEARD. (The HEARD moderator will vocalize your "Interview Greeting Message" and "Interview Ending Message" in the voice that you have chosen.)

Publish Your Study

  1. All other study settings, including information to collect from participants and recruitment should be filled out normally

  2. Preview your study to understand what your participants will experience

  3. Publish for your participants to begin interacting with your study

Question Tips for Voice Interviews

In your study design, we recommend the following to write questions:

  • Avoid complex sentences or jargon

  • Use conversational language

  • Test questions by reading them aloud

What Your Participants Experience with a Voice Interview

When participants join a voice interview, they'll go through these steps:

Modality Selection (if "Voice or Text" was chosen)

  • Participants choose either voice or text (This choice can be changed if the user chooses to restart the interview)

Consent and Permissions

  • Browser will request microphone and speaker permissions

  • Consent form with required checkbox agreement

  • Clear instructions provided if permissions are blocked

Welcome Message

  • HEARD introduces itself using your selected voice and set expectations for the conversation based on your Interview Greeting Message

During the Interview

Your participants will see clear status indicators. When they need to respond to a question, they will need to click on a button to begin recording and processing their answer, and click again when they are done.

  • AI is speaking - When the moderator is speaking or asking a question

  • AI is listening - When waiting for the participant's response

  • AI is thinking - When processing the participant's response

All "open-ended questions" will be asked via the AI. All structured questions, like multiple choice or ratings, will be asked via a visual interface that appears for manual selection. Participants will be asked to choose their answer and click "Next"

Questions with media, including images, videos or prototypes will display visually.

Voice Insights

You will be able to access the same filtering and insights tools from voice interviews as text interviews. However, if your participants respond via voice, you will also be able to access audio playback of the interviews.

Transcript View

You can access each user's individual transcript by going through the Participant Table and clicking on the Interview ID. Alternatively, in the Question Analysis tab, a side-out comes out when clicking on an individual insight. Each quote is attributed to a user and you can get to the transcript view by clicking on the user's name.

From the transcript view, you can click on the audio playback to Listen to complete interview recordings. Jump to specific moments in audio by clicking on that moment in the text transcript.

Participant Table

In your participant table, you will see a column labeled, "Playback". If the participants had a choice between voice and text, here you can see what they chose. If they chose voice, you will be able to click on a playback to listen back to their whole interview.


Voice interviews create more engaging experiences and often generate richer, more detailed responses from your participants. Need help? Write to us at tjehan@elis.ai if you experience any issues or have questions.

Did this answer your question?