How Greenfields Research compressed three weeks of mixed-method fieldwork into 24 hours - using Yazi's WhatsApp AI interviewer to reach young people across the Western Cape for the City of Cape Town.
.png)
How Greenfields Research compressed three weeks of mixed-method fieldwork into 24 hours — using Yazi's WhatsApp AI interviewer to reach young people across the Western Cape for the City of Cape Town.
Greenfields Research is a South African research house with deep roots in quantitative work — CATI (computer-assisted telephone interviewing), email-based surveys, and structured questionnaire design. When the City of Cape Town commissioned a mixed-method study into the experiences of young people finding work in the Western Cape, Greenfields needed a way to layer authentic qualitative depth onto their proven quant methodology, at pace.
Partnering with Yazi, Greenfields deployed a WhatsApp-native AI interviewer that ran 56 dynamic, conversational interviews across 36 Western Cape suburbs — all in a single afternoon. What would traditionally have taken three weeks of in-person fieldwork was completed in 24 hours, with comparable trend data and richer qualitative detail than the team had previously been able to capture at this speed.
Mixed-method research demands two things that rarely coexist: the structure of quantitative data and the texture of qualitative interviews. Traditional approaches force a trade-off — fast but shallow, or rich but slow.
Door-to-door fieldwork, venue hire, interviewer logistics, and geographic travel typically absorbed three weeks of project time before any insight could be drawn.
Physical fieldwork biases the sample toward accessible suburbs. Reaching young people across the full Western Cape — from Khayelitsha to Kraaifontein to Strand — required prohibitive travel.
Email surveys and cold outreach compete with every other notification. Traditional digital methods fail to hold younger audiences long enough to capture depth.
Classic questionnaires generate tick-box data. Focus groups generate rich language but at small scale. Neither, on its own, answered the City's question fast enough to be useful.
Yazi's platform gave Greenfields a conversational research layer that runs entirely inside WhatsApp. No app to download, no link to a web form — just a familiar chat interface where the AI interviewer asks, listens, and adapts.
The study combined WhatsApp's universal reach with an AI moderator capable of dynamic follow-ups and automatic voice-note transcription. Together, these reduced field time by an order of magnitude while maintaining qualitative texture.
Young people in the Western Cape received a WhatsApp message from Yazi Researcher Bot — the same channel they use to talk to family and friends — and began a natural, conversational interview. They could reply in their own words or record a voice note, which Yazi transcribed automatically for the research team.
The study deployed on the morning of 5 December 2024. By that evening, 56 completed mixed-method interviews had been conducted across 36 suburbs of the Western Cape — covering young people with formal jobs, those without, and those running side hustles. Trends aligned with Greenfields' established quantitative benchmarks, while the qualitative depth exceeded what traditional methods had been able to capture at comparable speed.
Where direct comparisons were possible, the WhatsApp AI-moderated trends mirrored the patterns Greenfields saw in their traditional quantitative work — validating Yazi as a rapid deployment tool that doesn't sacrifice rigour. The added value came from the why behind each data point: unprompted detail that structured questionnaires would never surface.
When asked open-ended questions about work, aspiration, and barriers, young people in the Western Cape answered with a clarity and specificity that traditional surveys rarely surface.
"Work means doing something you either love or something you need to do to earn an income."
"Be willing to start at the bottom, work for a short time for free and prove your worth."
"I have kids I need to provide for."
"Sometimes you need the money to start your own entrepreneurship — which I don't have."
"Don't always require them to have qualifications — many have skills not taught in schools."
"It means getting a salary and being able to provide for your family."
"Almost everything is digital now… there are a lot of scammers, but there's also genuine jobs."
"I'd ask for the government to provide free transport to all those who go out all day looking for jobs."
Greenfields accessed the study data through three purpose-built views inside the Yazi platform — full conversation transcripts, an auto-generated media library of voice notes, and themed insight clusters ready for client debrief.
| Traditional Mixed-Method | Yazi on WhatsApp | |
|---|---|---|
| Field time | ~3 weeks (door-to-door + venues) | 24 hours — single-day fielding |
| Geographic reach | Limited by field team travel | 36 Western Cape suburbs, zero travel |
| Qualitative depth | Focus groups, small sample | Dynamic AI probing at n = 56 |
| Quant rigour | CATI + email surveys | Structured cuts inside every interview |
| Voice data | Manual transcription lag | Auto-transcribed in real time |
| Venue & logistics cost | High — venue hire, travel, interviewers | None — platform only |
| Use cases | Full project only | Rapid scoping, main study, or both |
A new capability layered onto Greenfields' established quantitative practice — one that maintains methodological rigour while dramatically reducing field cycles. Yazi gave the team a rapid deployment tool suitable for category understanding, preliminary scoping ahead of larger studies, and full end-to-end research engagements alike.
Direct, unmediated input from 56 young people across 36 Western Cape suburbs — in their own words, in the language they use with family and friends. Evidence ready in days rather than months, enabling faster iteration on youth employment policy, training investment, and entrepreneurial support programmes.
To have a platform that can receive audio and then transcribe the audio in real time was quite exciting… If you can get people to open up, you've got a better chance of getting richer viewpoints from them.
Mixed methodology used to mean stitching together tools that weren't built to work together — a quant instrument here, a qual moderator there, a transcriber somewhere else, and weeks of coordination in between. This project demonstrates a different model: a single conversational platform, built for WhatsApp, that handles structured questions, open-ended probing, voice capture, and automatic analysis in one pass.
For research houses whose clients are asking for both speed and depth — government, brand, policy — Yazi collapses the field cycle without collapsing the method. Three weeks become 24 hours. One venue becomes 36 suburbs. One focus group becomes 56 individual interviews. The rigour stays; the lag goes.
Ready to run your next mixed-method study on WhatsApp?
Book a Demo →
%202.png)