6 EXECUTIVE SUMMARY The Department of Defense (DOD) has been engaged in a multi-year effort to modernize software technologies across its administrative (non-warfighting) and mission (direct warfighting) functions, and most recently, has accelerated its commitment to “acquire, deliver, and iterate on our weapon and business systems – including software – at speed and scale for our Warfighter.”1 As part of this whole-of-enterprise approach to modernization, integrating AI tools presents a critical opportunity to enhance, streamline, and improve technologies, processes, and support functions that serve the Warfighter. Collectively, the integration of AI technologies into the DOD represents one of the largest-ever migrations of technology into a single enterprise, and specifically, the Generative AI family of tools presents use-cases that impact nearly every part of the organization. We define this subset of technologies within the broader AI category to encompass Large Language Models (LLMs), Large Reasoning Models (LRMs), Multi-Modal Models and enabled or adjacent technologies including AI Agents and Retrieval Augmented Generation (RAG). Aggregating data sources from across the US federal government, academic leaders in Generative AI, and industry publications, we estimate that this family of technologies has the potential to improve labor-hour productivity across the Department of Defense’s active service and civil service by as much as 17% given the technology’s current abilities, impacting a wide range of enterprise-focused and mission- focused activities across the organization and serving the mandate for improving effectiveness across the force. However, while the Department of Defense has begun to develop, acquire and integrate these tools in the organization, the need for a holistic, replicable, and adaptable framework exists to independently assess the integration of Gen-AI tools across multiple layers of the enterprise to optimize return on investment, minimize technical debt, and ensure lasting and adaptable technological literacy and functionality. To address this need, our study establishes a framework, best practices, and considerations for the integration of Gen-AI tools into the DOD. Our framework includes: 1. Evaluating the organization for sufficient modernization to integrate Gen-AI tools; 2. Conducting a Relative Value Assessment of functional processes embedded within the military’s many divisions and teams, including an assessment of variables such as Cognitive Load, Impact, KPI Relevance, and Risk (each discussed in the study); 3. Planning multiple aspects of the necessary technological architecture of the organization against Gen-AI tools and identifying best practices in comparing potential vendors and models; 4. Executing the integration of the tools such that their value to the enterprise is preserved and expanded long-term through practices related to change management, and workforce training. The study then discusses notable considerations and dual-use applications that the private sector should monitor and apply to their own challenges with Gen-AI tools as the long-term integration of Gen-AI into the US military unfolds. Generative AI Adoption in the US Military