AI-Enabled Agentic Ecosystem for Enhanced Supply Chain and Procurement Decision Making

This project proposes an AI-driven agentic ecosystem to revolutionize supply chain analytics within the Air Force and Space Force. By integrating curator, data-retrieval, and synthesis agents with structured and unstructured data sources, the toolchain will scale data analysis at a speed unobtainable with traditional teams of data analysts. As a result, the project will make it possible to proactively manage the supply chain from end to end, ultimately enabling resilient and robust industrial base support for the Air Force and Space Force in both peacetime and wartime scenarios.  Curator agents will be in charge of (1) extracting knowledge from disparate structured data sources,  then (2) resolving entities within existing knowledge graphs to avoid potential duplicates while adding new knowledge, and (3) retrieving knowledge from open source indicators to add the type of intricate knowledge of the supply chain; necessary for meaningful analysis. Data-retrieval agents will be in charge of, (2) understanding user prompts, (3) selecting the best means to respond and prompting synthesizing agents if new information is required, and (3) retrieving knowledge from the graphs for the user using retrieval-augmented generation (RAG) techniques. Finally, synthesizing agents will evaluate new structured and unstructured data sources and integrate them for better insight.The agentic ecosystem will leverage and integrate with the Air Force Research Laboratory's Earth 616 program's secure, scalable architecture with decentralized graph databases and blockchain infrastructure. Through integration, this project will provide a Defense Industrial Base knowledge map, complete with intricate and nuanced relationships, which are impactful to the behavior of the supply chain, but virtually invisible with traditional data analytics methodologies.The project's success will be measured by its ability to reduce sole reliance on a finite number of data analytics professionals, empower them with scalability provided by an agentic ecosystem to integrate data sources, speed up response times to supply chain challenges, and expand understanding of the supply chain. This transformative tool promises to meet the immediate needs of the Air Force and Space Force, with broader commercial applications across various sectors. This STTR Phase I project aims to deliver a detailed plan and methodology for implementing an AI-driven agentic ecosystem that will be executed in Phase II. The expected outcome in Phase I is a proof of concept that demonstrates the AI agentic workflow’s ability to interpret complex data and add it to a knowledge graph. It will also provide a user interface that reduces the technical acumen required to retrieve supply chain insights. This technology has broader applications beyond defense, with use cases in commercial sectors requiring complex data analysis and decision support, of which there are many.