How Culture Foundry used Yazi's AI Interviewer on WhatsApp to deliver in-depth qualitative research at scale — preserving the probing, immersive methodology they're known for, while cutting roughly 20 hours off every project.
.png)
How Culture Foundry used Yazi's AI Interviewer on WhatsApp to deliver in-depth qualitative research at scale — preserving the probing, immersive methodology they're known for, while cutting roughly 20 hours off every project.
Culture Foundry is a market research agency renowned for rich, deeply immersive qualitative insight — the kind of work usually reserved for a moderator, a microphone, and an hour of undivided attention. Their clients include leading consumer brands looking for the "why" behind the "what".
Yazi partnered with Culture Foundry as an ongoing research partner, deploying its AI Interviewer via WhatsApp across multiple studies for different end clients. The AI was trained to ask — and probe — in Culture Foundry's signature voice, freeing the team from the logistics of manual fieldwork without losing the depth their reputation is built on. One exemplar project — an in-depth consumer study for a leading global beverage brand — delivered 50 high-quality interviews at pace and is used throughout this case study to illustrate the approach.
Culture Foundry's manual, in-depth interview methodology produced exceptional insight — but the process was difficult to scale under real-world deadlines.
Every manual in-depth interview demanded recruiter and moderator time, stretching studies over days and limiting the number of participants per project.
Tight client deadlines often forced a compromise: fewer participants, shorter interviews, or less probing — each of which eroded the quality Culture Foundry was known for.
Tracking who had and hadn't finished an interview was a manual, spreadsheet-bound task — making it hard to spot drop-off or plan follow-ups in real time.
Would an AI interviewer really probe like a human? Culture Foundry was understandably cautious about handing fieldwork to a machine — the depth of insight was non-negotiable.
Yazi introduced its AI Interviewer, designed to replicate the depth of a human interview with the speed of a digital survey. The interviewer was tuned to Culture Foundry's probing style — asking follow-ups, seeking examples, and giving participants space to answer in voice notes when text wouldn't do the emotion justice.
Participants are invited via WhatsApp. They answer a 20+ question guide at their own pace, choosing between voice notes and text. The AI probes, clarifies, and keeps the conversation moving — all without a human in the loop.
Participants receive an invitation via WhatsApp, tap to start, and answer a long-form interview in the chat interface they already use with friends and family — replying in voice notes or text, on their own time.
On the exemplar project, Culture Foundry completed 50 high-quality in-depth interviews in a fraction of the time their manual process would have required — saving approximately 20 hours of researcher time on a single study, while maintaining the richness of insight their clients expect. That same time saving has repeated across subsequent Culture Foundry studies run on Yazi.
A selection of participant responses from the exemplar study — the kind of texture and specificity Culture Foundry's methodology is designed to surface.
Culture Foundry accessed every in-flight study through the Yazi dashboard — a live completions tracker, full interview transcripts, and the underlying voice and text responses, all in one place.
| Participant | Progress | Status |
|---|---|---|
| Participant 01 | Full | Completed |
| Participant 02 | Full | Completed |
| Participant 03 | Full | Completed |
| Participant 04 | Part | In progress |
| Participant 05 | Part | In progress |
| Participant 06 | — | Not started |
| Manual IDI Process | Yazi AI Interviewer on WhatsApp | |
|---|---|---|
| Moderator time | One interviewer per participant | AI runs all interviews in parallel |
| Fieldwork duration | Days to weeks of scheduling | ~20 hours saved per study |
| Response format | Text-heavy, transcript after the fact | Voice & text, auto-transcribed |
| Scheduling | Calendar coordination per respondent | Participants reply on their own time |
| Completion tracking | Manual spreadsheet | Live real-time tracker |
| Depth of probing | Moderator-led follow-ups | AI probes tuned to Culture Foundry's style |
A scalable IDI methodology that protects the quality of insight the agency is known for — fewer hours on fieldwork logistics, more hours on analysis and storytelling. A live completions view turned status chasing into a glance.
Faster turnaround on in-depth qualitative research, with voice-note richness intact. More participants per project without a proportional increase in cost or timeline — giving brand teams a stronger evidence base for decisions.
Initially, we were struggling to track who had completed and who hadn't from our database of participants. Yazi built a table that shows who has completed and who hasn't, allowing us to track progress in real-time — which was a game-changer for us. Yazi's AI Interviewer exceeded our expectations, delivering the depth and quality of insights we require, but in a much more efficient manner.
One participant even commented, "The interviewer seemed to know so much about me that I kept on chatting to it even though I didn't need to anymore."
The Culture Foundry partnership is a proof point for what AI-led qualitative research can be when it's held to the standard of a world-class moderator. By meeting participants on WhatsApp, automating the fieldwork, and giving researchers a live view of completion, Yazi turned a methodology that was difficult to scale into one that travels well across studies and clients.
For agencies whose competitive edge is the depth of their insight, the lesson is clear: AI doesn't have to mean lower quality. With the right probing and the right platform, it can mean more of it — delivered in hours, not days.
Ready to run your next in-depth research project on WhatsApp?
Book a Demo →
%202.png)