Intelligent National Map: A Vision for Distributed and Agentic Geospatial Intelligence

Samantha T. Arundel, Wenwen Li, Kevin McKeehan, Bryan B. Campbell, Jung-Kuan Liu, Lawrence V. Stanislawski, Ethan J Sharvers, Greg Matthews, Philip Thiem, Dalia Varanka, Lynn Usery

An Intelligent National Map (INM) can change how worldwide mapping agencies, such as the U.S. Geological Survey, deliver the geospatial foundation of the Nation, as well as the capacity for the public to engage and use the data. It is envisioned as an innovative system that can coordinate analysis for spatial questions using structured reasoning grounded in semantic relationships, domain rules, and validated workflows. It is designed to coordinate analytic workflows autonomously and deliver validated, relevant geospatial answers and data quickly. This paper presents the technical and conceptual foundations of an INM, centered on multi-agent orchestration, natural language interfaces, retrieval-augmented generation, and access to trusted public data. Unlike static systems, an INM initiates from user input, which it analyzes and then activates task-specific agents to locate, analyze, and explain results in a transparent and accountable system. The platform represents a paradigm shift toward autonomous geospatial systems that could be capable of internal review, self-correction, and continuous learning. An INM can open new possibilities for changing access to geospatial intelligence, improving scientific rigor and decision-making, and building adaptive infrastructure aligned with federal science integrity principles.