This essay discusses and provides examples of a new type of assessment that involves computer-to-human conversations aiming to capture rich evidence of learning not unlike a conversation between a tutor and a student. Conversation-based assessment (CBA) builds on research from educational learning games, simulation-based training environments, and intelligent tutoring systems. The focus is, however, on assessment. A benefit of using automated computer-based systems for CBA is that they provide an intelligent, interactive, and open assessment capable for use in large-scale applications. This is the 25th issue in the R&D Connections series.