Greg Burke
Published on:
Jun 3, 2024
For years, qualitative research has had a scale problem. Great interviews take time to conduct, transcribe and analyse. And when things get busy (as they always do), it’s tempting to fall back on typed surveys. They are shorter and easier, but often lack depth.
That trade-off is starting to disappear with the surge of conversational AI.
With voice-first research in particular, we’re seeing a shift: from typed answers that feel like forms to spoken ones that feel like conversations. And what’s emerging from those conversations is richer, more spontaneous, and more emotionally honest than anything we’ve seen from traditional surveys.
In short, voice is giving us better answers.
A 2024 study from the IZA Institute of Labor Economics analysed over 7,700 open-ended responses to a single question. Some were typed and some were spoken.
The findings? Striking, but not surprising.
People who answered out loud gave longer, more personal, and more spontaneous responses. Participants shared stories, emotions, and reflections, not just bullet points. Even when their language was simpler, what they said held more weight.
And it wasn’t just a “feeling”, language models were used to evaluate the informativeness of each response, and the spoken ones consistently came out on top.
As Tobias Dengel put it in a Forbes (2024) interview, the shift to voice isn’t just about convenience, it’s a fundamental change in how we use technology. We speak faster than we type (up to five times faster than typing on mobile - hence we say tiny keyboards lead to tiny responses), and when machines can understand and respond contextually, everything becomes more intuitive and natural.
He also makes a crucial point about access. Voice-first systems make complex tools usable for a much broader group, especially those who may find typing slow or frustrating. As interfaces evolve, Dengel believes voice will become the default way we interact with machines. And we’re inclined to agree.
We’ve seen this firsthand many times. In one recent project spanning six countries and over 1,800 voice interviews, spoken responses were more than three times longer than typed ones. But it’s not just about volume, it’s about what you hear in a voice: the nuance, tone, hesitation, and emotion. Those non-verbal cues are something a written comment just can’t capture, but they come through clearly when someone’s speaking.
We get it. Voice may feel like a leap for some teams. But the barriers are lower than ever, and the returns are real.
Even a handful of voice responses can reshape how you think about your audience. You’ll start to spot not just what people say, but what they mean. And the best part? You don’t need to be a tech expert to use voice-first tools, they are easy to run, and the insights are clearer than ever.
You just need a tool that handles the heavy lifting, so you can focus on what matters: the stories, the patterns, the insight.
Voice-first research isn’t just a smarter way to understand people, it’s a tool every team should have in their research tool arsenal.
When you make space for people to speak, you get more than data, build trust, spark connection, and open the door to deeper insight and better decisions.
As voice technology becomes more intuitive and analysis more accessible, the teams that embrace it now will stay ahead.
I’m Greg and I’m the co-founder of a new kind of research platform called Tellet. We use AI to conduct and analyse consumer research interviews for faster, deeper and more affordable insights.
Want a free trial? Book a demo with us, or drop me an email – greg@tellet.ai.
