We often judge ability based on what people have done - their CVs, years of experience, outputs, and so on. But these metrics are often shaped by opportunity, not potential, and don't tell us the full story. What if we looked at something deeper than ability: how people think? In this talk, I explore how interaction can reveal patterns of thinking that are otherwise invisible. From using machine learning to analyse the data from a puzzle-based digital game, to building interactive game experiences that let people influence real-world outcomes, we begin to see signals of problem-solving, exploration, hesitation, and decision-making emerge. Often, these signals are more meaningful than traditional measures of ability - whether you're evaluating people or designing systems for them to use. Under the hood, this comes down to how we design systems: what we capture, what we ignore, and how we interpret user behaviour. I'll share practical ways to think about interaction data; not just as events to log, but as signals to understand. This talk bridges game design, frontend engineering, interaction design, and behavioural data - showing how we can move beyond tracking what users do, to understanding how they think.