Appropriate more funds. If this is an issue of citizen stakeholder engagement, I ask someone point in the necessary direction besides my Congressional reps.
Government's ability to attract people who could execute a project like this requires different compensation and career incentives. Base pay is capped at <$150k for the highest GS level at the highest step. There's also no real potential for bonuses or equity.
Beyond pay, government careers fundamentally optimize for low-risk decision-making. The goal is to not get fired over 20 years so that you can retire with a pension. This is why contractors like BAH gets hired: you, as a government program manager, don't get fired for going with a brand name even if they fail. If you hire some unknown development firm with great tech skills and they fail, you get canned.
There's also a lack of bold leadership and urgency that is customer-experience focused. Healtchare.gov benefitted from some amazing engineers, but the true catalyst for its comeback was that Obama realized it was a do-or-die initiative for his administration. His team moved heaven-and-earth to steamroll entrenched vendors, recruit talent, and hold people accountable.
Leadership and talent are what make the difference.
https://www.usds.gov/report-to-congress/2016/
It's a little better than it used to be with FedRAMP and such. But even now, agencies are still relying on broken-by-design contracts for terrible proprietary software.
To be fair, it's a really tough problem to solve. It would be a tough problem even without the help of the entrenched federal IT contractors (who would only stand to lose a huge amount of business).