Qualitative research has a scale problem: the depth comes from conversation, and conversation doesn’t scale. Sayle is our answer — an AI interviewer that ran 200+ Indonesian-language interviews across a studio’s release slate in twelve weeks. Here is what it does well, and where it still needs a human in the room.
“A survey tells you what. Only a conversation tells you why.”
Surveys scale and interviews don’t. That’s the trade every research team quietly makes, and it’s why most Gen Z insight work stops at a percentage. We built Sayle, our own AI interviewer, because we refused to make it — and because the Warner Bros. engagement needed 200+ qualitative interviews conducted in Indonesian, on a twelve-week clock, with limited local bandwidth on the ground.
What Sayle actually does:
Asks the follow-up: It hears a thin answer and pushes on it, the way a good moderator would, instead of marching to the next line of the guide.
Works in the respondent’s language: We ran 200+ interviews in Indonesian, because nuance does not survive being translated in someone’s head before they answer.
Runs at panel scale: Hundreds of real conversations in the window a human team would need for a dozen.
Stays consistent: Every respondent meets the same core structure, so transcripts from different regions are actually comparable.
What we still do ourselves:
Writing the guide and deciding what is worth asking
Handpicking respondents out of our 900+ student panel
Reading every transcript, not just the summary
Coding the themes and arguing about what they mean
Turning findings into recommendations a client can act on
Sayle is an instrument, not a research team. It removes the bottleneck that used to force a choice between depth and volume, and it leaves the judgment where it belongs — with people who read this audience well because they are this audience.
Koy Strategy.
Insights team.
Research.
AI.
Tech.
