How Inicio Insights ran a three-day pasta ethnography across 31 Nigerian kitchens using AI-moderated WhatsApp interviews - with 97% completion and zero manual coordination.
How Inicio Insights used AI-moderated WhatsApp ethnography to run a three-day pasta consumer study in Nigeria — replacing manual diary collection with structured, multimedia-rich fieldwork.
Inicio Insights, an independent market research agency based in Nigeria, was commissioned by a major food manufacturer to run consumer ethnography for a pasta product in development. The end client wanted more than a focus group — they needed to see pasta in the kitchen: how Nigerian households actually buy it, cook it, season it, and serve it.
The scope called for a three-day diary that captured personal context on day one, dry product evaluation on day two, and the live cooking experience on day three. Each household needed to upload photos and voice notes across the full preparation journey — not just describe it after the fact.
Inicio had run WhatsApp diary studies before, but the logistics were punishing. Collecting responses "one after the other" across 100+ participants, then stitching together voice notes, photos and text into something analysable, was — in their own words — "really hectic".
Consumer ethnography over multiple days is one of the most operationally demanding qualitative methods. It asks a lot of participants, and traditionally it asks even more of the research team.
Running a WhatsApp diary "one message at a time" across 30+ participants, three days in a row, with follow-ups on every interesting answer. A team-intensive task before a single insight surfaces.
Photos of shopping baskets, voice notes about texture, videos of cooking — scattered across individual chats with no structure. Analysis couldn't start until media was pulled together manually.
A moderator can probe deeply in a focus group of ten. At diary scale, every participant gets the same scripted follow-ups — the interesting answers don't get chased the way they would face to face.
Once the fieldwork ended, the real work began — transcribing voice notes, tagging media, and cross-referencing answers across three days and multiple households before the end client could see anything.
Yazi designed a three-day AI-moderated diary delivered entirely through WhatsApp. Each day had a distinct objective, and the AI interviewer adapted its follow-ups to each participant's cooking style, household context, and pasta habits.
Each day opened on WhatsApp with a themed set of prompts, probing media uploads (photos, voice notes, video), and adaptive follow-ups generated by Yazi's AI based on the participant's own answers.
Nigerian households don't need a new app to participate in research — WhatsApp is already where they live. The AI interviewer met them there, in a conversational format that felt more like chatting with a curious friend than completing a survey.
Three-day diaries usually bleed participants — drop-off compounds across waves. Inicio's pasta study held steady: 31 households started on Day 1, and 30 completed through to Day 3.
The AI interviewer didn't stop at the scripted question. Where a participant mentioned a specific preparation ritual — pasta prepared in "room temperature water" rather than boiling, for instance — the AI probed further, turning a one-line answer into a detailed account of why, when, and for whom.
Inicio's researchers moved between three core views: a Media Library for the cooking photos, videos and voice notes; per-participant Transcripts with the AI's adaptive follow-ups; and a Participant Journey that stitched the three days together as a single timeline.
| Traditional WhatsApp Diary | Yazi AI-Moderated Diary | |
|---|---|---|
| Coordination | Moderator sends each prompt manually | Themed questions run automatically across three days |
| Probing | Same follow-ups for every participant | AI adapts follow-ups to each individual's answer |
| Media | Scattered across individual chats | Auto-indexed by participant, day and question |
| Voice notes | Transcribed later, by hand | Auto-transcribed and aligned to the question |
| Client visibility | End of fieldwork at earliest | Live dashboard, same-day |
| Scale | Limited by team capacity | 30+ households managed from one platform |
Inicio received a structured, ready-to-analyse data package designed to fit both their own workflow and the end client's reporting needs.
The end client — a major Nigerian food manufacturer — received qualitative depth plus structured quantitative reads on preference drivers, all from the same study. They came back asking for numerical representation of preferences and immediately requested Yazi on a follow-on project, specifying it by name.
Inicio ran a richer three-day ethnography with a fraction of the coordination overhead they'd previously needed. The end client's response — asking to use Yazi on future work — turned one project into a repeatable capability, and positioned Inicio as an innovation leader in Nigerian consumer ethnography.
I'm happy to tell you that the client recommended you. They said "this is what we want to use" going forward. One of the big reasons was speed — the results came in faster than with traditional methods. What's funny is my colleagues, even the competitive ones, were asking "what did you do to this guy? He loves this platform!" He even asked us to switch a different project onto Yazi.
Consumer ethnography has always had to choose between depth and scale. The moment you add households, days and multimedia, the cost of running a rich study traditionally rises faster than the quality of the insight. That trade-off is what Inicio's pasta study stopped being bound by.
By meeting participants on WhatsApp — the platform they already use — with an AI interviewer that probes as deeply as a good moderator would, Inicio ran a three-day ethnographic diary that captured the real kitchen, with the real voices, and delivered it to the end client fast enough to shape product development. The end client noticed. Then they asked for it again.
Interested in running a similar study?
Book a Demo →
%202.png)