Google Public Sector launches Gemini to 3 million Department of War personnel on GenAI.mil

➡️ Subscribe to our newsletter for the latest insights: https://goo.gle/4thTbM6
➡️ Explore our thought leadership hub: https://publicsector.google/ai/

Executive summary: Google Public Sector, in partnership with the Office of the Under Secretary of War for Research and Engineering helped the Chief Digital and Artificial Intelligence Office (CDAO) build and launch GenAI.mil in 54 days, giving 3 million department personnel a secure, enterprise-grade platform to support their daily operations and workflows.The deployment has scaled to 1.5 million+ active users, trained more than 55,000 personnel during rollout, and enabled the workforce to build over 100,000 custom domain-specific AI agents. In one early deployment, a custom agent helped clear a 27,000-document policy review backlog in weeks instead of an estimated 11 years.

Challenge: To protect the nation and maintain mission readiness across diverse defense environments, defense leaders required fast, secure access to operational data. Historically, software adoption faced fragmented silos, leaving hundreds of high-potential AI pilots disconnected from the personnel who needed them most. Solution: In partnership with Google Public Sector, CDAO launched Gemini Enterprise on GenAI.mil in just 54 days—30 days ahead of schedule. Built on a secure-by-design foundation, GenAI.mil delivers high-throughput scale while adhering to strict defense security guardrails. Powered by Gemini Enterprise, the environment gives defense personnel an intuitive agentic stack to build, orchestrate, and govern mission-specific AI agents across every branch.

Results: GenAI.mil serves as the first enterprise AI capability across the ecosystem, giving over 3 million personnel access to cutting-edge commercial technology, including same-day deployment of Gemini models upon commercial release. The platform has scaled to over 1.5 million active users, supported live interactive training for 55,000+ personnel, and empowered teams to create 100,000+ custom domain-specific agents. Using these custom agents, one team resolved a 27,000-document policy review backlog that was estimated to take 11 years manually in just a few weeks.

Key takeaways and highlights from our interview with Deya Banisakher, Acting Director, Frontier AI at CDAO:

→ “Bottom line at CDAO, we turn the Department's data into decision advantage for our warfighters and our workforce. Primarily, we are the hub for data analytics and AI... AI-ready platforms that are enterprise-grade, secure, and trusted.”
→ “The department is so vast and large and innovative that we had hundreds of pilots that were being built and dying on the vine... the data was there, the pilots were there, the users were there. It's just that they weren't finding each other.”
→ “We are no longer in an era where we adopt a finalized, finished product off the shelf... We're moving at the speed of commercial AI, and that’s where the Secretary wanted us to be, to bring decision advantage for our warfighter, and that is to move at the speed of the mission.”
→ “What it took to get an enterprise platform for 3 million plus users within an unheard-of speed and matter of days—54 days, to be specific... was a lean team that operated at speed in a startup culture and was able to deliver on the direction that we got from the Secretary.”
→ “The single biggest lesson and insight I've gained... is to fail fast and fail big, and then recover it. To scale quickly and not look for a perfect solution is to look for the doers around me... move fast, deliver a platform, start small, but build for enterprise grade.”

Google Cloud products used: Gemini Enterprise, Google Cloud Storage, Google Kubernetes Engine (GKE)