Making evidence the engine of continuous digital delivery

Imagine a federal product team that’s six months into modernizing a public-facing app. Early testing of the legacy app uncovers one workflow that users consistently struggle with. The research team discusses the findings with developers in a kickoff briefing. Then, the modernization moves into build mode while the testing data moves into a folder.

Months and several sprints later, the same user issue pops up in a test of the nearly finished product. At this point, fixing the problem doesn’t mean a quick tweak. It requires a redesign.

This expensive challenge is common and has a simple cause: research and other valuable data that aren’t built into a development team’s ongoing workflow tend to fall out of use. Nielsen Norman Group, which studies user research in agile teams, has found that research that gets left out of a team’s task list gets set aside while research built into the backlog continues informing decisions, sprint after sprint.

This failure to keep information in reach can have ramifications across an entire modernization effort, not just in one team’s task list. Different teams are generating research, testing, and performance data throughout the effort. What they often lack is a reliable way to share what they learn with each other, or to find it later when they need it.

Prioritizing global access to those insights allows them to shape decisions before they’re made instead of after the fact. Teams that take this approach benefit from reduced friction: fewer redundancies, fewer decisions based on stale evidence, and less time bringing one side of a project up to speed on what the other side already knows.

Three federal agencies demonstrate what this approach looks like in practice through team planning meetings, shared dashboards, and AI-assisted rapid prototyping for user testing.

A standing seat in sprint planning

In one federal agency, the UX research team sat in on weekly meetings where the delivery team planned its work on a public-facing digital platform. Instead of routing findings through a handed-off report, researchers regularly provided evidence-based insights directly to developers deciding what to build next. In one case, user research showed that a single term meant different things to different people. Because that finding reached developers at nearly the same time, eight separate teams were able to start from a common understanding quickly.

Thanks to continuous insights, the development team made more than 15 content and terminology changes to its website in just over six months, and the site’s use doubled. This demonstrates how continuous access to fresh data doesn’t just keep projects running—it can help teams deliver better results. These benefits scale up when the audience isn’t just one delivery team but an entire program.

One view of all the evidence

Another agency asked the question: What would change if every piece of a program’s evidence—what’s working, what isn’t, and what’s already been tested—were visible to everyone at once? The agency’s answer was a real-time dashboard that brought all of it together. Status-reporting effort dropped by more than half, and leaders gained a live, evidence-based reference for prioritizing projects.

By connecting strategy, delivery, and measurable outcomes in a single view, the dashboard helped leadership allocate resources with more confidence and spend less time “discovering” what was already known somewhere else in the program. This capability is enhanced when the dashboard receives new information at a fast clip.

Research at the speed of delivery

A third agency sought to shorten the time lag between when information is collected and when it’s used, essentially creating a constant testing process. Using AI, the agency was able to build rapid prototypes of digital health tools and test them with real users. This testing took days instead of months and required about one-tenth the effort of previous approaches. Test findings were continuously available to designers and developers rather than arriving in a single batch at the end of a phase.

By applying AI across research, design, and development, the agency eliminated the "hand-off tax"—the time and context lost each time work passed from one person to the next. The agency was able to compress months of sequential work into a just few days of steady, shared progress.

Conduct research that powers the work

Although contextually different, each of these examples demonstrates that continuously generating new information and ensuring it reaches teams in real time is how organizations turn data into value. When treated as one-time deliverables, research, project metrics, and user data can’t achieve their full potential. But when built into the daily rhythm of a program and made accessible to those who need them, these insights function like a shared, evolving utility rather than a static report. They keep every decision—from a single sprint to a major budget request—grounded in what’s true right now, not what was true several weeks ago.

This shift from producing information to keeping it in circulation is what turns scattered research, metric, and testing results into an engine for sustainable federal delivery.