The OECD.AI Policy Navigator

ANALIA automates the extraction, processing, and structuring of data from handwritten and digital PDF documents submitted to OSITRAN as part of the road infrastructure maintenance supervision process. The initiative enables automatic data capture, integrates it into a predefined structure, applies technical criteria automatically, and generates information for regulatory decision-making, reducing manual tasks and improving operational efficiency in the supervision process.

Konsekvenshjälpen is an AI assistant developed to help public officials produce higher-quality regulatory impact assessments. It provides step-by-step guidance based on Sweden’s Impact Assessment Ordinance and the Agency’s official guidance, helping users explore alternatives, identify affected stakeholders, and improve drafts among other tasks.

This AI platform enabled all French citizens living outside France to become better informed about the complex issues related to Social Protection for French Citizens Abroad. This approach marks a major democratic innovation, making a technical subject accessible, ensuring an inclusive debate and laying the foundations for more informed and constructive citizen participation at the international level.

Agroboost is a national programme that helps farmers and local producers use artificial intelligence and connected devices to improve productivity and decision-making. It was created to address low productivity, climate risks, and limited access to technology in rural areas. The initiative benefits small and medium agricultural producers by strengthening skills, increasing efficiency, and supporting more sustainable farming practices.

The LINC initiative is a national network of community innovation labs that provide hands-on training in digital technologies such as robotics, 3D printing, and basic artificial intelligence. It benefits students, teachers, entrepreneurs, and small businesses across different regions of Costa Rica. The initiative was created to reduce gaps in access to technology and strengthen digital skills needed for future jobs and local economic growth.

A pathway developed by the Municipality of Bologna to govern the use of AI responsibly by combining strategic intent with operational delivery. It adopts a hybrid model in which critical decisions remain centralised, while certain capabilities may also be activated externally, with the aim of progressively strengthening key internal competences over time (e.g., risk management). All of this follows a single process, from ideation through to monitoring.

The policy sets Rio de Janeiro State’s guidelines for strengthening the AI ecosystem, improving public data management, and promoting the ethical use of data and AI in government. It defines responsibilities, instruments, the AI Forum, and the Ethics Council for AI so this public policy can be implemented, monitored, and continuously improved through expected results, indicators, and targets. It can benefit citizens and give public servants and managers greater legal certainty when using AI.

JOSEIAS is an AI-powered assistant designed to support municipal civil servants in navigating the Electronic Information System (SEI). It provides 24/7 real-time guidance on administrative processes, reducing technical bottlenecks and improving the efficiency of public management. The initiative helps bridge the digital skills gap among government employees using generative AI.

The JULIO AI Platform is Peru's leading public sector digital transformation project. It uses AI to optimise registry services across three pillars: Specialised Consultation (natural language processing and large language models for legal precedents), a Learning Centre (manuals, prompt libraries, and automated "esquela" generators), and Integrated AI Access (specialised tools for registry contexts). Rather than replacing human judgment, it empowers staff to improve legal precision and efficiency

Galileu is an artificial intelligence tool developed by the Regional Labor Court of the 4th Region (TRT-RS), Brazil, that assists judges in drafting decisions by reading case files and organising key issues and arguments. It automatically generates structured summaries and suggested text based on case data and a controlled legal database, with full judicial review and editing. The initiative aims to improve efficiency, accuracy and timeliness in judicial workflows.

Fala.BR is the single-entry platform for ombudsman submissions to Brazil’s Federal Government. However, fewer than 3% of submissions are linked to a specific public service, limiting service evaluation from the user’s perspective. This initiative uses artificial intelligence to automatically connect citizen submissions to relevant public services, helping institutions improve public services and their quality and effectiveness based on citizens’ experiences.

Trending Topics in Ombudsman Submissions uses artificial intelligence to identify common themes in submissions sent by citizens to the Brazilian Federal Government through the Fala.BR ombudsman platform. It shows how these topics change over time and highlights sudden increases in complaints or concerns. This helps public institutions respond faster and coordinate actions across agencies.

Our initiative uses artificial intelligence to help digitise historical Serbian texts. Many old books and documents are difficult for modern software to read because of complex layouts and decorative elements, which leads to poor digital quality, especially for smaller languages. By improving how these texts are processed, we support libraries, archives, museums, researchers, and the wider public in preserving and accessing Serbia’s cultural heritage.

Bergen is developing an AI-enabled curbside analytics tool combining CDS/MDS-based geofencing with privacy-preserving video AI. Blurred video streams classify vehicles and street users within digitally defined zones to measure usage, conflicts, and capacity. The goal is evidence-based, adaptive curb management—improving safety, logistics, micromobility integration, and shared space planning near the city square.

This internal initiative empowers government employees to propose and test small-scale AI proofs of concept to improve their daily administrative tasks. It was developed to build internal expertise and speed up processes, effectively removing the reliance on expensive external vendors. By upskilling staff and defining clear internal AI roles, the organisation aims to become entirely self-sufficient in developing and evaluating its own AI projects.

The guideline on generative artificial intelligence was created to support and encourage civil servants to use AI solutions.

The AI Accelerator connects state agencies with vendors to rapidly prototype AI solutions. It provides a structured platform for agencies to submit unresolved challenges and collaborate with vendors on rapid prototyping through 60-day sprints. This programme enhances operational efficiency, decision-making and problem-solving for agencies by fostering collaboration to leverage advanced AI technologies tailored to their specific needs.

This national strategy aims to transform Slovakia into an energy-computational hub by leveraging carbon-free energy for AI Factories. It addresses labour shortages and fiscal constraints by automating public administration and creating a sovereign AI adaptation layer. The initiative centralises AI governance to ensure data sovereignty, ethical standards and economic growth in line with EU Digital Decade 2030 goals.

This project adapts CANChat, a secure federal generative AI tool, into a customisable resource for HICC. Models are being fine-tuned, and internal knowledge repositories are being built to streamline departmental workflows. The project is built with accessibility in mind, ensuring that these powerful AI insights are inclusive and fully available to all staff who need them.

It applies ML techniques to identify and flag anomalous patterns in large volumes of bridge sensor data from HICC-owned major bridges. The initiative focuses on improving data quality by detecting irregular, missing or inconsistent sensor readings, thereby producing cleaner datasets. These cleaned datasets can be used by engineers and analysts to support monitoring activities related to structural integrity, infrastructure safety and long-term asset management, including lifespan extension.