Skip to main content
Generate a link and send it to respondents over a digital channel, such as email or WhatsApp, or however you reach them. Respondents complete the interview on their own device, in their own time. This method needs a reliable connection, a device, and enough digital comfort to finish unassisted. This is not always realistic, for example where there is poor connectivity or low digital literacy in the population you are studying. You can still use voi, with a person in the loop instead of the respondent going it alone. In practice, whoever is running the study opens the same interview link and acts as the point of contact:
  • In person, on the ground: sit with the respondent. Either hand them the device so they answer directly in their own voice, or read the questions aloud and enter their answers on their behalf if they would rather not handle the device. If you read aloud, read each question exactly as written and one at a time, so every respondent is asked the same thing.
  • Over a phone call: call the respondent, and open the interview on your end. Either type in what they say as you go, or put the call on speaker so voi’s microphone picks up their voice directly.
In either case, the real interview is still running underneath. The AI still asks live follow-up questions based on what the respondent just said. voi analyses the response the same way as one collected through a self-service link. Only the person conducting the interview needs a connection and a device. The respondent never does.
Coming soon: voi will be able to place the phone call itself, and ask the respondent the interview questions directly. No person will need to make the call.

​
Decide this before you launch

Match the collection method to your population before you send a study out, not after a low response rate tells you the self-service link isn’t working. If poor connectivity or low digital literacy is likely in the group you’re studying, plan for a person-in-the-loop method from the start. Whichever method you use, the welcome screen states that an AI system is conducting the interview before the first question, and the default consent text repeats it. Keep that statement in place if you customise the text, since EU transparency rules require respondents to be told. See Share for the full walkthrough of sending a study out, and Data privacy for what respondents are told about how their data is used.