Expanding AskGDR to Better Serve an Evolving Geothermal Industry

Abstract

The U.S. Department of Energy's (DOE) geothermal AI research assistant, AskGDR, has been expanded to better meet the needs of its users and the geothermal industry. Originally the result of integration with the metadata and supporting documents from data submissions to DOE's Geothermal Data Repository (GDR) and a Large Language Model (LLM), AskGDR now includes industry reports, content from DOE's GeoBridge, and the past 9 years of proceedings from the Stanford Geothermal Workshop. The National Laboratory of the Rockies (NLR), in response to user feedback and analysis, has expanded the corpus of knowledge included in AskGDR to better answer the questions asked by users. AskGDR now provides answers to users’ questions about the geothermal industry at large, emerging technologies and trends, and cutting-edge research, while still enabling users to interrogate the deeper aspects of GDR data and the methods used to derive them. This paper outlines the expanded corpus of knowledge input into AskGDR, an analysis of questions asked, the efficacy of recent improvements, and impact and quality of the answers generated.