Human-AI Tools for Transforming Researchers & Innovators

Experts across domains often fail to solve problems not due to lack of knowledge, but because implicit paradigms in their field constrain how and what solutions are produced. In this project, we're developing a structured LLM-based system that guides experts through problem definition, assumption surfacing, and solution ideation to support innovation across a variety of fields. The core idea behind this approach is that LLMs can act as facilitators for reflection and ideation, helping experts distinguish between challengeable field norms and legitimacy criteria, enabling them to question dominant paradigms while maintaining defensibility within their field. Our findings so far show that LLM-assisted workflows enable greater breadth and depth of idea generation, support deeper reflection on implicit assumptions, and help produce solutions that both overcome the limitations of existing approaches and remain legible to the field. This opens up endless possibilities for problems to be solved in practice, from scientific research to product development and policymaking.