Better government services start with a clear use case.
Explore where Claude could help your team, choose the right way to use it, and build an honest case for cost and public value.
The planning path 01 — 04
01 Define the use case Start with the work and who it serves.
02 Choose a Claude route Chat, Claude tasks (Cowork), Code, or API.
03 Model the cost Make volume and assumptions visible.
04 Test public value Measure quality, capacity, and service.
Start with the mission Show the assumptions Validate before scaling
01 / Find the fit
What does your team need to accomplish? Select a starting point to see a possible Claude route. Recommendations are illustrative and should be validated against your agency's requirements.
01 Find answers in policy guidance Help staff research and draft with source material. ↗ 02 Prepare recurring casework Turn files and notes into reviewed deliverables. ↗ 03 Modernize an internal service Help a technical team build and maintain software. ↗ 04 Support residents at scale Bring AI into an agency-owned digital service. ↗
A possible starting route 01 / 04
Claude Chat
For staff who need a thinking partner to research, summarize, and draft while they retain responsibility for the final answer.
First pilot Test with a small set of approved policy documents and measure time to a verified answer.
Validate before use Confirm access controls, permitted data, source review, and human approval for consequential responses. See matching use cases ↗
Illustrative recommendation, not a security approval or product eligibility determination.
02 / Choose the route
Four ways to put Claude to work The best route depends on who does the work, where it happens, and whether the agency is building its own application.
01 / Individual work Claude Chat Research, reason, write, and analyze with a person directing the work.
02 / Knowledge work Claude tasks Delegate bounded work across files and tools with the Cowork capability in Claude, then review the deliverable. Availability varies by plan.
03 / Software teams Claude Code Support developers as they understand, build, test, and maintain software.
04 / Agency applications Claude API Build Claude into a service or workflow your organization owns and operates.
Government availability · checked September 30, 2026 Anthropic announced Claude Code and Cowork in public beta for Claude for Government Desktop in a FedRAMP High authorized environment. This planning map does not establish an agency's access, data permissions, or authorization for a workflow. Read Anthropic's announcement ↗
Federal and defense procurement context A September 25, 2026 appellate ruling upheld an exclusion of Anthropic from the defense department's supply chain. It does not determine every civilian agency's access. Federal and defense teams should confirm current procurement status and applicable restrictions before planning a deployment. Read the court opinion ↗
03 / Make the case
A useful estimate shows its work. Cost alone cannot tell you whether a service is better. Build a pilot case around workload volume, full staff effort, operating cost, capacity, service measures, and evidence from a bounded test.
Build a pilot value case ↗ 01 Define the baseline How is the work done today, and what does success mean for residents and staff?
02 Model cost transparently Use current, dated inputs and the right cost method for each Claude route.
03 Test the value Measure quality, time returned, adoption, and service outcomes before scaling.
