How The Research Agency (TRA) used AI-moderated diary studies to uncover in-the-moment betting decisions across Australia and New Zealand.
How The Research Agency (TRA) used AI-moderated diary studies to uncover in-the-moment betting decisions across Australia and New Zealand.
The Research Agency (TRA), a leading qualitative research firm based in Auckland and Sydney, needed to understand the real drivers behind sports betting behaviour. Their client wanted more than survey data — they wanted to see what actually happens in the moments when someone places a bet.
Traditional approaches — retrospective interviews, online surveys, focus groups — would only capture what participants remembered and were willing to share days or weeks later. The behaviours TRA needed to study are fleeting, emotional, and deeply contextual: a bet placed in a pub during a live game feels very different from one placed alone on a couch.
TRA needed a method that could follow participants through their natural betting routines over multiple weeks, capturing decisions as they happened — the environment, the emotion, the trigger, and the outcome.
Sports betting is a behaviour that resists traditional research methods. The decisions are fast, emotional, and heavily influenced by context that disappears the moment it passes.
Asking bettors to describe their decisions after the fact produces rationalised narratives, not authentic accounts. The emotional context — the adrenaline, the social pressure, the impulse — is lost.
Where someone bets, who they're with, what they're watching, and what promotions they've seen all shape the decision. None of this is captured by a survey form completed at a desk.
Betting isn't a single event — it's a pattern shaped by wins, losses, moods, and routines across days and weeks. A one-off interview captures a snapshot, not the full picture.
Running identical methodology across New Zealand and Australia — with different betting cultures, platforms, and regulations — required a flexible, scalable approach.
Yazi designed a multi-day digital diary study delivered entirely through WhatsApp, combining timed broadcast check-ins with adaptive AI-moderated conversations. The study ran in two phases: a pilot in New Zealand (18 participants) followed by the main study in Australia (26 participants).
Three campaign waves timed to intercept participants during peak betting windows — not when it was convenient for the research team, but when bets were actually being placed.
Participants received timed prompts via WhatsApp during peak betting windows, then engaged in AI-moderated conversations — sharing text responses and voice notes capturing their in-the-moment decisions.
Across the 14-day study, participants remained actively engaged — responding to timed broadcasts and continuing AI-moderated depth conversations across all three waves. The conversational WhatsApp format kept drop-off low and data quality high.
The AI interviewer adapted its questioning based on each participant's responses — probing deeper on betting routines, triggers, and the role of different platforms. Participants engaged naturally, sharing voice notes and detailed text responses that revealed patterns traditional surveys would miss.
TRA received a comprehensive, analysis-ready data package — individual transcripts, structured exports, and rich media organised by participant and wave. The platform's built-in views made it easy to explore patterns directly.
| Traditional Qual | Yazi Diary Study | |
|---|---|---|
| Timing | Retrospective recall, days or weeks later | In-the-moment capture during live events |
| Context | Described verbally from memory | Voice notes and descriptions of the actual environment |
| Depth | Fixed question sets, one-size-fits-all | AI-adapted follow-ups based on individual responses |
| Media | Text only | Text, voice, screenshots, video |
| Duration | Single session or interview | 14-day longitudinal study across multiple touchpoints |
| Scale | Separate field teams per market | Same platform across NZ and AU, centrally managed |
TRA received a comprehensive, analysis-ready data package designed to integrate with their existing qualitative workflows.
Authentic, in-the-moment insights into betting behaviour that retrospective methods simply cannot produce. Evidence of the environments, triggers, and emotional states that drive betting decisions — giving the end client rich, contextual data to inform product and messaging strategy.
A methodology that delivered ethnographic depth at diary-study scale, across two markets, without proportional increase in field team or operational overhead. The NZ pilot allowed TRA to refine the approach before the Australian rollout — a built-in quality control loop that traditional methods rarely afford.
We needed to capture what happens in the moment someone places a bet, not what they remember a week later. The AI interviewer picked up on the small details like a mention of mates texting or a Friday afternoon ritual, and probed further in a way that felt natural. We ended up with the depth of a one-on-one interview across 44 participants, and voice notes where you could actually hear the pub in the background. That's ethnographic context you don't get from a survey.
This project demonstrates that the traditional trade-off between depth and scale doesn't have to apply. By meeting participants on a platform they already use, at the moments that matter, through an AI interviewer that adapts to each individual, the study captured data that's not just more convenient to collect — it's fundamentally more authentic.
For research agencies working on complex behavioural studies, the combination of WhatsApp delivery, AI-moderated depth interviews, and rich media capture opens up territory that traditional diary studies and retrospective interviews cannot reach.
Interested in running a similar study?
Book a Demo →
%202.png)