Skip to main content
An audience model (we call them population models) is an AI representation of a specific group of people. It’s built from real survey data and can predict how that group would answer questions they’ve never been asked. Think of it as a research panel that’s always available and responds in seconds.
So far 49 public models and growing. UK and US adults, 16 UK Consumer Finance segments (from Growing Families to Just About Managing), tech industry segments (developers, designers, engineering managers, data scientists, and more), and B2B software buyers. Public models are free to use.
Custom models are built from your data: survey responses specific to your audience, your market, your customers. They’re private to your account and optimised for the questions that matter to your domain. Create one →
A CSV with survey questions, answer options, and response distributions. Minimum 4 questions, ideally 20+. The more relevant and varied the seed data, the better the model generalises to new questions. Full requirements →
Segments are distinct population models representing subgroups within a broader audience. For example, our UK Consumer Finance collection has 16 segments, each with different demographics, financial behaviours, and attitudes. You can ask the same question to multiple segments and compare how they respond.
Yes. As you add more seed data to a population, the model’s coverage and accuracy improve. More questions means more context for the model to draw on when predicting new ones. Accuracy is re-evaluated automatically after each update.
Every prediction comes with an accuracy estimate. If the model has low confidence for a particular type of question, the accuracy score will reflect that. We recommend using test batches to evaluate whether a model is reliable for your specific use case before relying on its predictions.