--- title: "About Alvar Laigna" description: "Alvar Laigna’s software and AI experience, published work and independent proposal for agentic government: bounded delegation with accountable public services." keywords: "Alvar Laigna, AI architect, independent architecture, Estonia, agentic government, agentic state, agentic gov, govtech, gov tech, AI agents in government, public services" canonical: https://agenticgov.eu/about url: https://agenticgov.eu/about --- # Alvar Laigna . Software and AI architect. Two decades across software development, machine learning, AI and electronics. I value spec-driven engineering where action precedes the blueprint. A martial discipline, uncompromising impact honed by lean, kinetic execution. [Experience and project history](https://alvarlaigna.com/about) ## Why this exists **agenticgov.eu** is an independent architecture for public-sector leaders and engineers: how citizens can delegate bounded actions to AI agents while public services retain final authority. Agents could make public services easier to use: by voice, on the move, through a personal assistant or a humanoid helping with everyday tasks. People should not need to learn the internal structure of government to get things done. That future needs a precise answer to a simple question: who authorised this action? This site brings together my writing on AI readiness, spec-driven engineering and public-service delegation. The argument can be read in full, followed through an architecture, and tested against a small, deterministic service boundary. The goal is practical, proactive help with public services, while people retain control and public authorities remain responsible for the decisions they make. ## A legal foundation Estonia’s Administrative Procedure Act gains **§ 7¹, “automaatne haldusmenetlus”** (automated administrative procedure). The amendment enters into force on **1 January 2027**; the new provision applies to procedures initiated on or after **1 July 2027**. This provides a relevant legal foundation for the discussion. It is not blanket authority for an agent to act: a citizen’s delegation and an authority’s automated administrative procedure remain distinct. The law preserves conditions for discretionary decisions and the right to personal contact and to be heard. [Read the amendment at Riigi Teataja](https://www.riigiteataja.ee/et/akt/109072026012). This English summary is unofficial. ## A European framework The **EU AI Act** adds a risk-based framework for AI systems across Europe. For high-risk uses, it sets requirements for risk management, traceability and human oversight. Transparency duties make certain uses of AI visible to the people affected. Requirements apply in phases; not every AI system has the same obligations. My architectural reading: a trustworthy service needs both a valid mandate and controls suited to its risks. Compliance with AI rules does not itself confer authority to perform a public task. [The European Commission’s AI Act guide and current implementation timeline](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai). ## Estonia’s AI direction Estonia’s **Data and AI White Paper 2024–2030** and **AI action plan 2024–2026** set the national policy direction. The [official AI roadmap](https://www.kratid.ee/kratt-visioon) brings those documents together. The [Eesti.ai initiative](https://www.riigikantselei.ee/en/eestiai-initiative) builds on existing work, developing AI skills and practical adoption across business and the public sector. The connection I draw is simple: Estonia’s ambition sits within Europe’s safeguards. Turn that ambition into better services through clear mandates, capable institutions and measurable public value. A roadmap guides change; it is not a legal permission to act. Policy context checked on 8 September 2026. These are architectural observations, not legal advice. ## Sources and permissions The essays, documents and summaries reproduced here are my own. Third-party standards and legislation are linked to their original sources. The architecture is a draft derived from the main essay; it is not presented as an official state specification or an existing national deployment. This is independent work. The views expressed here are my own, not an institutional position. ## The writing behind it - [Technology does not create public valueThe blueprint: process, legal authority, bounded delegation and evidence.](https://alvarlaigna.com/posts/technology-does-not-create-public-value) - [Spec-driven AI engineeringSpecifications, deterministic controls and human ownership of the result.](https://alvarlaigna.com/posts/spec-driven-ai-engineering) - [The missing link in the agentic stateThe question that started this: how can a service verify an agent’s authority?](https://alvarlaigna.com/posts/the-missing-link-in-the-agentic-state-ai-identity-crisis-in-public-services) - [A flag is not a verdictEvidence and human review in procurement intelligence.](https://alvarlaigna.com/posts/a-flag-is-not-a-verdict-cartel-detection) - [AI readiness in the public sectorInstitutional readiness, voice interfaces and accountable public services.](https://alvarlaigna.com/posts/kuidas-ehitada-ai-valmisolekut-avalikus-sektoris) - [The AI-ready enterpriseSystems agents can discover and use within explicit boundaries.](https://alvarlaigna.com/posts/the-ai-ready-enterprise-preparing-for-autonomous-agents) - [AI where it belongs: Riigihanked.euA practical example of separating deterministic work from AI assistance.](https://alvarlaigna.com/posts/how-ai-is-fixing-estonian-public-procurement-riigihanked-eu) ## Continue the conversation. [alvarlaigna.com/contact](https://alvarlaigna.com/contact) [GitHub](https://github.com/alvarlaigna)