A Prompt Engineering-based Process to Build Proto-personas during Lean Inception

Product discovery approaches such as Lean Inception (LI) typically span five days (40 working hours). During LI, the participants create and refine proto-personas during four working hours to understand user needs. Proto-personas are preliminary, assumption-based representations of ideal users that guide initial design discussions. The accuracy of proto-personas generated in this context has been counterintuitive due to limited time for idea exploration and refinement, for example. There are approaches to building personas (e.g. data-driven, LLMs). However, there is a gap in exploring the use of prompt engineering and proto-persona strategies to support the Product Discovery approaches. Our research investigates the application of a prompt engineering-based approach to building protopersonas during LI. We report an exploratory case study where six participants used our approach to generate proto-personas in a given scenario. The impact of our approach positively influenced the outcome. Most proto-personas developed by our process better represented the target audience than those from LI, despite some inconsistencies. Our process was well accepted by participants and suggestions were made to improve the process. Our approach used an average of 11 minutes of working hours (SD 2.24 minutes), traditionally this time in LI is four hours.