Generative Artificial Intelligence (Gen AI) is rapidly transforming how individuals and teams develop ideas and solve problems, yet empirical evidence shows that its effects vary widely depending on how humans interact with the technology. We studied the topic by means of three field experiments involving a total of over 50 Italian professionals in a Gen AI laboratory on small-team creativity and performance. The study adopted a three-stage empirical structure in which teams generated innovative solutions for a topic across repeated rounds, progressively integrating AI support: (round 1) human‑only brainstorming, (round 2) Gen AI‑assisted ideation, and (round 3) Gen AI‑assisted ideation following a prompting lesson.