Using Synthetic Focus Groups to Close Research Gaps

You’ve spent months building it, a mobile therapy app designed to support people when they need it most. The features are polished. The content is trauma-informed. The user flow is clean. Before launch, you run a pilot focus group to test the experience—gather feedback, catch blind spots, and adjust where needed. But something still nags at you. Is the feedback in your pilot group honest, or just agreeable? Who’s missing from the room entirely? And what about the people who never download mental health apps at all—would they possibly consider downloading this app? Groupthink, selection bias, and drop-off rates are familiar problems with significant consequences. Whether you're testing a digital health tool, launching a climate-conscious product, or piloting a public policy campaign, the challenge stays the same: how do you gather real insight before it's too late to change course? 

Synthetic focus groups are stepping into that gap. With the help of large language models (LLMs), researchers can simulate diverse perspectives, pressure-test ideas, and explore blind spots that traditional methods tend to miss. Done well, they don’t replace humans, they help researchers hear more of them. Behavioral science helps make synthetic focus groups usable, responsible, and ready for the real world. Especially when synthetic tools promise faster timelines and broader reach, those principles help ensure we’re not just hearing more voices. They help ensure we’re hearing what matters.