How US Digital Response mobilised 4,000 tech volunteers to fix government infrastructure

Executive overview

Government digital infrastructure has been neglected for decades — unemployment systems, benefits portals, and data pipelines were not built for crisis-scale demand. When COVID-19 hit, states and counties faced an impossible gap between what they needed to deliver and what their systems could do.

Jen Pahlka and two other former US Deputy CTOs launched US Digital Response (USDR) in mid-March 2020 to bridge that gap — matching volunteer technologists, designers, and data scientists with government entities requesting urgent help, for free.

The fastest route to better government is people who know both tech and public service working inside the system, not around it.

Origins: Code for America and the case for civic tech

• In 2009, Pahlka founded Code for America from the observation that Web 2.0 practices — user-centred design, iteration, data — could transform how government served citizens.
• The inspiration was not crisis but opportunity: if these methods could elect a president, they could improve public administration.
• Local governments were resistant but not immovable; enough early adopters created a thread that eventually changed the fabric of government.
• A visit to the UK's Government Digital Service (GDS) — which consolidated 4,000 government websites into one usable portal — showed what was possible at the centre of government, not just the edges.
• That visit directly led to the creation of the US Digital Service (USDS) inside the White House, modelled on GDS.

How USDS worked: tour-of-duty model

• Technologists came in for defined stints — often two weeks of "discovery sprints" — rather than full government careers.
• Many who arrived for two weeks stayed for years once they saw the scale of impact available in government.
• Projects included the Veterans Benefit Management System and the E-Verify program at USCIS.

How USDR was built in a weekend

• On 15 March 2020, Pahlka got a call from DJ Patil (first US Chief Data Officer) and realised every level of government would be overwhelmed.
• A Google Form for volunteer sign-ups went live Monday; by that weekend, 1,000 qualified people had responded.
• A website with two forms — one for governments to request help, one for volunteers to offer it — launched within a week.
• USDR remained an all-volunteer effort with no formal organisational structure; its core team of ~40 was split evenly between government and tech-company backgrounds.

How matching works

• Governments submit a problem (not a solution request); the USDR team does an intake call to understand context and infrastructure.
• A job description is written and matched against a searchable database of 4,000+ volunteers by skill, availability, and domain.
• Volunteers sign an oath reminding them to work with humility inside government constraints — a deliberate culture-clash reducer.
• Example: a New Jersey data file needing reformatting was received one evening and returned the next morning — faster than any normal procurement cycle.

Two categories of government need

• Data for decision-making: automating manual data collation (e.g. Pennsylvania's 100+ daily hospital data points into a dashboard), building data pipelines, and supporting modelling for reopening decisions.
• Service delivery: standing up new benefits portals and eligibility tools in days rather than years. Pandemic unemployment assistance is the flagship example — states normally have years to implement new benefits; USDR compressed that to days.

The deeper infrastructure problem

• Government systems were not modernised because there is no ribbon-cutting ceremony for technology upgrades — the public doesn't hold officials accountable for back-end infrastructure.
• The crisis revealed that a new digital service can be stood up in a day under pressure; the question is why it takes years in peacetime.
• Investment must go into both technology and the human capacity — people and mandate — to maintain modern systems between crises.

Digital equity and mobile-first design

• Making government services digital can widen the digital divide if done badly (desktop-only, complex language, 212-question forms).
• Done well — mobile-first, plain language, minimal questions — it narrows the divide: a seven-minute SNAP application means a food bank worker can help an older person enrol on the spot.
• Accessibility improvements benefit both end users and the eligibility workers processing applications.
• City of San Rafael example: a Code for America fellow helped move all remaining paper-based city services online, removing the need for residents to visit offices during COVID.

Remote work and team trust

• USDR's 40-person core team operated fully remotely; Pahlka had met only 8 of them in person before working together.
• High trust formed through shared urgency and purpose, not physical co-location.
• Experience prompted a reassessment of how much commuting and travel is genuinely necessary.