Skip to main content

Use cases

AI standards and tools

Toolbox overview

Skills and training opportunities

Legal and procurement guidance

AI use cases for the European public sector

The Joint Research Centre (JRC) has identified three key application fields for AI in the public sector in its dedicated report, each covering a range of concrete use cases.

  • Improving internal administrative processes; 
  • Delivering more accessible and personalised public services;  
  • Supporting evidence-based policymaking  

National examples already demonstrate tangible results (see below).

To obtain additional AI use cases, you can access them from the Public Sector Tech Watch Cases Viewer. This section showcases a collection of cases from across Europe on the adoption of artificial intelligence, blockchain and other emerging technologies in the public sector.  

The AI Toolbox will also include the Public Sector AI and Interoperability Readiness Pathway (PAIR), a structured roadmap to help public administrations implement these use cases. PAIR, which will be released soon, will provide step-by-step guidance, decision points, user journeys and practical examples to help public administrations move from identifying a need to operating an AI-enabled service.  

Advancing AI adoption Cover

AI application fields and use cases in public administration

3

AI-enhanced administration

This application field focuses on AI solutions that reduce administrative burdens, improve internal efficiency and strengthening organisational resilience. It includes solutions that support internal processes such as document handling, anomaly detection and staff learning. The JRC has identified the following five national use cases:  

AI for text enhancement

AI can improve written content by enhancing clarity, grammar and style. Writing assistants analyse text and suggest corrections and style improvements, reducing drafting time and improving consistency across administrative documents.

Example: Open-source GenAI for public administrations

The GenAI solution F13 serves as a German alternative to commercial LLM tools designed for public administrations. The Baden-Württemberg public administration innovation lab introduced this generative AI text assistance for state administration to reduce administrative workload. It offers features including text compression, cabinet document summarisation, a research assistant and text generation. Since 2024, all staff with access to the state administration network have been able to use the solution. 

Country: Germany | Reference: PSTW-2195

AI for document and data processing

AI can improve document and data management by automating tasks such as data extraction, classification and routing. Public administrations can use AI to extract relevant information from structured and unstructured documents, classify content and route it to the appropriate department, reducing manual workload and errors. 

Example: AI for thematic document classification

The website of the Official Barcelona Provincial Gazette uses the AI-based solution Cidobot to automate summaries. Cidobot employs neural networks to classify publications in the official gazette by thematic or procedural area. The solution improves classification and summary writing, helping increase efficiency and transparency. 

Country: Spain | Reference: PSTW-2169

AI for detection

AI can analyse data in real time to identify and report unusual events or conditions. This capability is relevant in areas such as cybersecurity, infrastructure monitoring and regulatory enforcement. AI inspection solutions can analyse location-based images and generate alerts, supporting more precise monitoring and greater coverage. 

Example: AI for spotting undeclared swimming pools

The French tax office implemented an AI-driven computer vision system to verify whether homeowners have declared their swimming pools. This led to the identification of over 20,000 previously undeclared pools, resulting in an additional EUR 10 million in tax revenue. The software automatically detects pools in aerial photographs and cross-references findings with real estate and tax databases. 

Country: France | Reference: PSTW-887

AI for recommendations and decision-making

AI can process complex datasets to generate recommendations that support human decision-making. It can support prioritisation of actions, allocation of resources and optimising services, increasing the consistency, precision and speed of decisions across governmental functions. 

Example: AI for smart public employment services

EMi (Emprego Intelixente) is an AI-powered tool deployed across all employment offices of Galicia’s public employment service. The solution combines data analytics, labour market intelligence and AI to build competency-based profiles, generating personalised recommendations that support career counsellors. The tool continuously improves by learning from anonymised user interactions, while ensuring that final decisions remain with human professionals.

Country: Spain | Reference: PSTW-2167

AI for learning

AI can support internal staff by enhancing job performance and facilitating continuous learning. Virtual assistants can give staff quick access to internal policies, procedural guides, and intelligent question-and-answer solutions which can address queries in real time, providing personalised support. 

Example: AI for improving access to public documentation

The Berlin Senate Chancellery and City LAB Berlin developed Parla, an AI assistant that generates response suggestions by extracting information from over 11 000 files held on the parliamentary documentation website. The objective was to streamline research using the Berlin public administration’s documentation by minimising manual search.

Country: Germany | Reference: PSTW-2252

2

AI-enabled people-centric public services

This application field focuses on AI solutions that improve interactions between public administrations, citizens, businesses and other stakeholders, making public services more accessible and user-friendly. The JRC has identified the following four national use cases:  

AI for service accessibility

AI can improve public service accessibility by removing language and communication barriers. Automated translation services, voice interfaces and simplified access tools can make services more inclusive by removing linguistic, sensory or cognitive barriers. 

Example: AI for social claims

CLAIM is an AI solution that helps users anonymously determine what state benefits they are eligible for and provides personalised advice. The solution simplifies access to state support by calculating individual claims based on a single form. AI, including machine learning and natural language processing, powers the solution’s interactive forms, dynamic data extraction and multilingual capabilities.

Country: Germany | Reference: PSTW-2398

AI for personalised and proactive services

AI can enhance public services by analysing individual needs and data to deliver personalised and anticipatory experiences. AI algorithms can assess a user’s past interactions and preferences to suggest relevant services or proactively send reminders for upcoming deadlines.

Example: AI for proactive administrative procedures

A pilot programme in Catalonia explores the use of AI in proactive public services: automatically applying annual property or water tax discounts without requiring citizen action, and pre-filling applications for school lunch aid. The goal is to direct public resources to those in genuine need, rather than those who are simply adept at navigating bureaucracy.

Country: Spain | Reference: PSTW-1071

AI for participation and co-creation

AI can empower citizens to actively participate in the design and improvement of public services and policies by contributing ideas or feedback. AI-driven platforms can collect and process large volumes of citizen input, automatically clustering similar suggestions, identifying recurring themes and summarising findings for decision-makers.

Example: AI for participatory urban co-design

The City of Helsinki used the UrbanistAI platform to involve residents in co-designing street transformations. During participatory workshops, citizens uploaded photographs and contributed ideas. The platform used generative AI to produce photorealistic visualisations based on participant inputs. The two co-designed streets were subsequently built as proposed.

Country: Finland | Reference: PSTW-2001

AI for service engagement

AI-driven channels such as chatbots and recommendation engines can enhance user engagement by keeping individuals informed and involved throughout their interactions with public services. This use case supports ongoing interaction, clarifies next steps and offers timely prompts that aid users in completing procedures.

Example: AI for answering questions about public competitions

A collaboration between Formez (the Italian public administration modernisation body) and CSI Piemonte produced Formez’s first virtual assistant. Camilla, the digital assistant, uses generative AI to answer questions about competitions managed by Formez, operating around the clock to reduce waiting times and help citizens navigate public competitions more easily.

Country: Italy | Reference: PSTW-2334

1

AI-assisted policymaking

This application field focuses on AI solutions that support analysis, generate useful recommendations and facilitate evidence-based policymaking processes. The JRC has identified the following four national use cases:  

AI for policy assessment and alignment

AI can help assess whether existing or draft policies comply with legal and institutional frameworks. It can check syntax and semantics, detect overlaps or contradictions and evaluate consistency with national or supranational legislation. These capabilities also support 'law as code' initiatives by making legislation easier for machines to interpret.  

Example: AI for supporting the legislative process

The GenAI4Lex-B project introduced three prototypes aimed at enhancing understanding of regulatory sources, streamlining the amendment process, and checking compliance with existing regulatory frameworks. The project utilises a hybrid and interdisciplinary AI approach, focusing on effective collaboration between AI and humans and on ensuring technological quality, repeatability, explainability and validity of results.

Country: Italy | Reference: PSTW-2194

AI for trend analysis

AI can support the analysis of historical and current data to detect emerging patterns and forecast future developments relevant to governance. It can anticipate demographic shifts, economic fluctuations or environmental risks, supporting forward-looking policy design and resource planning. 

Example: AI for forecasting demand for emergency medical services

The LifeSaver project, led by the Estonian Health Board, enhances emergency medical services by improving resource allocation, optimising service areas and using AI-driven predictive analytics to forecast system demands. Funded by the 2021–2027 EU cohesion policy, LifeSaver seeks to enhance emergency care and strengthen public sector innovation capacity.

Country: Estonia | Reference: PSTW-2352

AI for evidence-based policy-drafting

AI can support the creation of new public policies based on empirical evidence and data-driven recommendations. Predictive models can help identify key issues, understand social and economic dynamics and assess potential policy impacts, strengthening the quality and transparency of policy formulation.

Example: AI for supporting legislative drafting

In a pilot funded by Sitra, the Finnish Innovation Fund, the Ministry of Transport and Communications tested generative AI based on Finnish large language models to support legislative drafting. The pilot produced a chatbot interface through which legal drafters could search Finnish legislation and obtain direct links to relevant provisions. The tool was tested in the context of the national implementation of the EU Data Act.

Country: Finland | Reference: PSTW-2260

AI for data exploration

AI enables public administrations to analyse large and complex datasets to identify useful insights, revealing hidden knowledge across governmental functions such as public health, social assistance, and environmental and economic affairs. 

Example: AI for exploring and mitigating the impact of extreme weather events

The Norwegian Meteorological Institute and the European Centre for Medium-range Weather Forecasts are entrusted with the Destination Earth project, aimed at creating a highly accurate digital model of the earth. They are demonstrating how machine learning and deep learning can enhance datasets and refine climate and weather predictions to offer more accurate insights for mitigating extreme weather impacts.

Country: Norway | Reference: PSTW-2382