Dify
Menu

Solutions

Public Sector

AI for government.

One sovereign platform, self-hosted on home soil, that lets non-technical staff in any agency build their own apps, and lets the good ones be shared across government instead of each office rebuilding the same thing.

Contact Sales
Contact Sales

Deployed in Production at.

A major metropolitan government's digital agency

A digital agency set up by a major metropolitan government runs a shared generative AI platform on Dify Enterprise for the metropolitan government and its 62 wards and cities.

The platform is hosted domestically by the agency, following open-source-first and self-hosted principles. Each agency workspace is isolated, while a dedicated sharing layer turns proven apps into digital public resources that other municipalities can reuse.

Three categories are already in production: document-workflow efficiency, Q&A over internal data, and business-specific assistants.

Read the full storyHow a Major Metropolitan Government Runs Shared AI for 62 Districts on Dify Enterprise
Contact Sales

  • 62

    Wards and cities on one shared Dify platform

  • 3

    Categories of use case live on Dify

  • 14M

    Metropolitan residents served by agencies on shared Dify

  • 23+39

    Special wards + cities/towns co-building on one platform

Source: Dify Enterprise deployment at a major metropolitan government's digital agency

Why
This Industry, Why Now.

The constraints

Public sector organisations operate inside three hard constraints. Data sovereignty is non-negotiable — citizen data, policy data and operational data must stay within national boundaries and on infrastructure the public body controls.

Procurement and approval cycles are long, which means AI platforms need to work for many agencies rather than being rebuilt for each one.

And non-technical staff in agencies, bureaus and local government need to be able to use these tools without depending on central technical teams for every workflow they want to run.

The role of AI

For this industry, the platform question is not whether to adopt AI. It is whether one shared, sovereign environment can serve many agencies without requiring each one to stand up its own.

The organisations that answer this well run one Dify Enterprise instance centrally, operated by a digital agency or shared services body, with each agency getting its own workspace.

Patterns We See.

Sovereign, self-hosted deployment
Dify Enterprise runs on the public body's own infrastructure, inside its own jurisdiction, with no data leaving the perimeter. The deployment model matches what public sector data governance requirements demand — not adapted to fit, but designed to operate this way from the start.
One platform, a workspace per agency
A central body runs the shared platform. Each agency, bureau or city gets its own workspace, with its own data connections, its own workflows and its own access controls. The platform carries the infrastructure and governance. Agencies carry the applications relevant to their own operations.
Reuse across agencies, by design
Most public sector applications solve generic problems — document workflows, Q&A over internal policy, meeting minutes, constituent inquiry handling. Applications built by one agency can be adopted by others, rather than each agency rebuilding the same thing independently. The platform accumulates shared capability over time.

Discuss Public Sector Workflows with Dify.

Contact Sales