The Agentic State

Institutional framework · 2026

The Agentic
State

An institutional model of the state in the age of artificial intelligence

Whitepaper · Whitepaper v1 — coming soon
The state is no longer a body of rules.It is becoming the operating system a society runs on.

The thesis the whole framework rests on.

Framework

The architecture of the agentic state

Five load-bearing structures of the new model — from the computational core of the state to its programmable interfaces.

AI State Stack

The vertical architecture of state action: four layers on which the agentic state runs.

  1. L1

    Data layer

    A single, verifiable representation of reality: data on people, objects, processes and context.

  2. L2

    Decision layer

    Models and agents that weigh scenarios, risks and consequences for every decision.

  3. L3

    Execution layer

    Automated and semi-automated execution of decisions through public services.

  4. L4

    Interface layer

    Where citizens meet the state: agents, applications, APIs and channels of explanation.

01

Zero-Request Government

A state that acts proactively — granting services and payments without an application from the citizen.

02

AI Caseworker

AI takes on the first tier of bureaucracy: triaging cases, checking eligibility and preparing decisions.

03

Government App Store

A modular ecosystem of public services — installable, updatable and reusable.

04

State as API

Government as programmable infrastructure: public powers exposed as secure, auditable interfaces.

Initiatives

Ten initiatives of the agentic state

Applied directions in which the model becomes concrete systems.

  1. 01

    Zero-Request Government

    Proactive services and payments with no application required.

  2. 02

    Digital Citizen Twin

    A consistent digital representation of rights, statuses and interaction history with the state.

  3. 03

    AI Caseworker

    AI at the first tier of case handling and decision preparation.

  4. 04

    National Decision Engine

    A shared environment for modelling the scenarios and consequences of public decisions.

  5. 05

    AI Health System

    A shift from treating consequences to early diagnosis and prevention.

  6. 06

    AI Education Companion

    Personal learning agents and lifelong learning.

  7. 07

    Government App Store

    A catalogue of modular public services.

  8. 08

    AI Urban Operating System

    Mobility, infrastructure and safety managed as one urban system.

  9. 09

    Regulation Simulator

    Testing laws and reforms in a model before they reach citizens.

  10. 10

    State as API

    Programmable, auditable interfaces to public functions.

Research

Whitepaper and research access

The agentic-state framework is being formalised in a whitepaper and a series of publications. This is research access, not a store.

Whitepaper v1 — in preparation
Request access

The whitepaper and key materials are published in the project’s Telegram channel as they are released.

About

What this research is

The Agentic State is a research project on a new institutional model of the state, in which artificial intelligence becomes part of the basic architecture of government.

The aim is to describe how the state sees, calculates, decides, explains and remains accountable when part of its action is carried out by models and agents — and what limits, rights and lines of control this requires.

This is not a book project or a product. The book “The Agentic State” is only one artifact; the core of the project is the system of ideas, the whitepaper and further research.

Go to the book