The Algorithmic State Architecture

An integrated framework for AI-enabled government

Based on Engin, Crowcroft, Hand & Treleaven (2025)

Download the briefing (PDF)Read the Full Article

Overview

Governments are using artificial intelligence more and more: to deliver services, to build evidence for policy, and to support or automate decisions. In most places, however, these tools are introduced one area at a time, each with its own budget, rules and goals. This briefing explains why that piecemeal approach often holds digital programmes back, and what a more joined-up approach looks like.

It draws on the Algorithmic State Architecture (ASA), a framework set out by Zeynep Engin, Jon Crowcroft, David Hand and Philip Treleaven. ASA brings together four established fields, digital public infrastructure, data for policy, algorithmic government and GovTech, and treats them as interdependent layers of an emerging AI-enabled state. Each layer builds on and enables the others, forming a coherent whole that is greater than the sum of its parts.

Looking at the experience of Estonia, Singapore, India and the United Kingdom, the briefing finds that success depends less on excellence in any one layer than on how well the layers work together, and on governance developing in step with technology.

The four layers

Service layerGovTech

The apps and portals citizens use

Process layerAlgorithmic government

Where decisions are supported or automated

Intelligence layerData for policy

Turning data into evidence for policy

Foundation layerDigital public infrastructure

The shared digital systems: identity, payments and data exchange

Read from the bottom up: each layer enables the one above it, while the needs of the higher layers, in turn, shape the foundation.

Key messages

1The building blocks of AI-enabled government depend on each other. Treated as separate projects, each can work well on its own but fail to fit together.

2Every layer is limited by the others, in both directions. A weakness in one place can hold back the whole, so the real problem is often not where it first shows.

3Technical and governance choices go together. They work best when they are designed at the same time, not one after the other.

4A clear vision must come first. Technical choices should follow clear policy goals, ideally goals that last beyond a single political cycle.

5There is no single path. Estonia, Singapore, India and the United Kingdom took different routes. Local context should decide what is built first.

6Democratic oversight matters more as algorithms take on more authority. It should be built into each layer from the start, not added later.

What the briefing covers

1Why fragmented adoption holds governments back

2The four layers, and what makes each succeed or fail

3How the layers depend on one another

4Four national experiences

5Assessing maturity: a practical lens

6Recommendations for policymakers

7Conclusion

Who it is for

The briefing is written for policymakers, public servants and practitioners who shape or deliver digital transformation in government. It is a plain-language summary of the original framework paper, and it links the framework to the principle of systemic coherence set out in the DSA's working paper, Governing Well in the Algorithmic Age: The Foundations of Digital Statecraft.

Source and citation

Cite the Full Article
Engin, Z., Crowcroft, J., Hand, D. & Treleaven, P. (2025). The Algorithmic State Architecture (ASA): An Integrated Framework for AI-Enabled Government. Preprint. https://doi.org/10.48550/arXiv.2503.08725

Cite this briefing
The Digital Statecraft Academy (2026). The Algorithmic State Architecture (ASA): An integrated framework for AI-enabled government. Policy Briefing. https://zenodo.org/records/22816072

Join the conversation

We welcome comments on this briefing, as well as examples from those applying the framework in practice.

For inquiries: office@digitalstatecraft.academy