What Lies Beneath the Keystrokes

What Lies Beneath the Keystrokes: Unlocking Insights Through Typing Patterns in Conversational AI Research
David Wright & Karlien Kriegler, Hello Ara
Conversational AI has rapidly transformed the research industry, powered by Generative AI’s ability to enhance probing, automate verbatim coding, and create richer, more dynamic respondent experiences. As a result, more people are typing—and typing more often—in surveys than ever before.

But Conversational AI isn’t just about adding a couple of open-ended questions to the end of a closed survey. It’s a craft: built on thoughtful conversation design, deeper follow-up probing, and a fundamentally different interaction style. This change in interaction affects good respondents and fraudsters alike. Following an extensive academic review, we began keylogging select projects to uncover what typing itself can reveal. The findings have been fascinating.

In this session, we’ll explore and challenge the following hypotheses:

• Can typing patterns distinguish between System 1 (fast, intuitive) and System 2 (slow, deliberate) thinking?
• Can we tell when someone is truly engaged with the survey?
• Can fraud be identified without reading the text, just from the typing behaviour?