Scaling a secure Navy AI capability from 250 to 7,200+ users

Situation

OPNAV-GPT, an enterprise secure AI capability for the Navy, had working technology and 250 pilot users. Leadership needed it to become an operational tool that thousands of people could access, trust, and use in their daily work.

Constraints

• Two networks, NIPR and SIPR, each with separate access and approval processes
• Stakeholders spanning government leads, product owners, developers, data and machine-learning teams, platform administrators, security, and end users
• Users with widely varying familiarity with AI tools and clear limits on what the tool could be used for

What I did

As the delegated day-to-day delivery lead, I coordinated more than 10 cross-functional workstreams and kept priorities, readiness, and releases aligned with mission needs. In parallel, I built the adoption program:

• Separate onboarding paths for NIPR and SIPR, each matched to its access requirements
• Training and communications that set clear expectations for appropriate use
• A library of 50+ reusable prompts so new users started from proven examples
• Support readiness and feedback loops that routed user issues into the product backlog

Why it worked

Delivery and adoption were planned together from the start. Access, training, and support were treated as backlog items with owners, not as launch-week tasks.