Letter from leadership

We’re deeply familiar with hard challenges. We founded Nava in 2015, after our work to support the HealthCare.gov recovery effort, based on the insight that the gap between what government promises and what people receive is mostly a delivery gap — one that can be closed. A decade into this work, we’d frame it differently: government can only deliver when it has enough capacity and control. When an agency doesn’t own its data, codebases, interfaces, or priorities, government ends up renting its mission back from vendors incentivized to make themselves impossible to replace. When a government agency does own those things, many benefits follow. First, vendors have to compete on results, because the agency can replace them (including us!). Next, the agency can reuse what one team builds, so every engagement leaves government collectively more capable than before. And finally, no vendor can hoard insider knowledge of a system everyone else is locked out of, which keeps competition honest. That is what an open, accountable digital ecosystem looks like, and in 2026, Nava built toward it with more urgency than ever before.

This included putting AI to work carefully, responding to emergencies, and advancing structural change. Civil servants began gaining fluency with AI tools and agentic techniques, as the vendor hype cycle clicked into high gear. Our approach was to pursue concrete value from the jump, running pilots and scaling up what worked. At the Department of Veterans Affairs, we cut the manual work of classifying disability claims by 80%. In Pennsylvania, tools we shipped within five months of H.R. 1 becoming law cut illegible submissions by 80%, so fewer people lose food assistance over a paperwork error. And at Nava Labs, our philanthropically funded prototyping arm, we piloted an AI referral tool with Goodwill; two-thirds of the case managers surveyed said it made them more confident in their referrals. The same principle applies to complex legacy systems. In New Jersey, our AI analysis of 1.1 million lines of unemployment insurance code cut the time engineers needed to understand it by about 30%. At the Centers for Medicare & Medicaid Services, we worked with payment systems handling more than $1 billion in claims per day, moving their monitoring onto a more cost-effective platform with zero downtime. Making systems and infrastructure reliable and flexible to change is only part of the solution. The next critical step is helping agencies maintain sufficient control over these systems. Early in 2026, we acquired Beam and folded it into Strata, our open source architecture and toolkit for running modern government services. With Strata, we helped Minnesota launch a complete paid leave program, on time and under budget, that paid out nearly $600 million to 75,000 Minnesotans in its first six months — a program the state now controls. We kept building Simpler.Grants.gov with the U.S. Department of Health and Human Services in the open, where Simpler Search now handles more than 70% of Grants.gov search traffic, and anyone can read the code and contribute. And we began modernizing how the U.S. Department of State determines passport renewals. Our government’s digital ecosystem remains fragile, but the work in this report shows what’s possible when agencies increase their control and capacity to deliver. Veterans get decisions faster, Minnesotans receive paid leave on time, beneficiaries access Medicare securely, and grant seekers find opportunities more easily. We’re proud of how we delivered these human-centered outcomes at all levels of scale, in a broad spread of contexts, across varying degrees of urgency.

There’s still so much work ahead. The thriving digital ecosystem this country needs is one where agencies routinely deliver successful outcomes for people, and vendors make government more capable, not more dependent. That standard starts with us. We invite every agency to hold us to it, and every vendor to compete with us on it.