A field-level agenda for government that works

Government needs to work. AI gives us an opening

A generational technology shift is changing what small, capable teams can build and how quickly they can learn. The opportunity is not to bolt AI onto government as it exists. It is to use this moment to build the lasting capacity government needs to deliver.

The technology changed. The obligation did not.

01

Government must deliver

People experience government through whether its institutions and services actually work. Delivery capacity is not secondary to policy. It determines whether policy becomes real.

02

AI creates an opening

AI changes the economics of building software, analyzing information, and redesigning work. It gives government and its partners new room to move.

03

Capacity is the goal

The field should use this moment to change how government builds, learns, funds, buys, and owns, leaving behind institutions that can keep improving after the moment passes.

Seven priorities for the field

These are choices about where the field should concentrate effort, authority, talent, and money. They are a starting position, not a catalog of every useful activity.

  1. 01

    Make product delivery a core government capability

    Build permanent, multidisciplinary teams that own services, measure outcomes, release frequently, and improve continuously.

  2. 02

    Build permanent public talent and institutional capacity

    Give government the product, engineering, design, data, operations, and leadership capacity to remain an intelligent owner.

  3. 03

    Use AI to redesign work

    Reconsider whole services, processes, roles, and controls instead of using new tools to accelerate old routines.

  4. 04

    Build shared public digital infrastructure

    Invest in reusable components, standards, data infrastructure, and institutions that solve recurring problems across government.

  5. 05

    Align funding, procurement, and oversight with delivery

    Fund capable teams, learning, measurable outcomes, and responsible iteration rather than static plans and prescribed solutions.

  6. 06

    Turn implementation knowledge into field capacity

    Make practical experience, evidence, decisions, and reusable artifacts easier to find, trust, adapt, and improve.

  7. 07

    Build guardrails into delivery

    Make privacy, security, accessibility, and accountability part of product decisions, architecture, testing, and operations.

A strategy should move resources and behavior.

The agenda will connect priorities to concrete commitments from organizations and leaders across the field. A commitment has an owner, a time horizon, and a measure of progress.

No commitments are being claimed yet. The Workbench now includes candidate commitments: concrete bets we are exploring and scoping before anyone has agreed to own them.

See the candidate commitments →

This is a point of view to build on and argue with.

The public agenda will stay clean and coherent. The Workbench shows what is in development, where input is useful, and eventually what happened to the ideas people contributed.

You will be able to

  • Challenge the diagnosis
  • Share evidence and implementation experience
  • Improve or propose an action
  • Make or support a commitment

Contributions will inform editorial decisions. Recoding America will retain final responsibility for what appears in the Agenda.

Workbench now in development →