A Closer Look at Generative AI Pilots in Spain’s Public Administration

There is much discussion about artificial intelligence (AI) and how it will affect our lives, jobs, and relationships. Public administration and citizen services are also areas where its influence will be felt. A study by AI Watch, the European Commission's knowledge service, indicates that since 2020, Spain has been implementing an AI adoption strategy that includes initiatives such as training programs for public sector workers.

Beyond these strategic plans, is generative AI already being used in the provision of public services in Spain? Are there solutions with this technology already being implemented to improve citizen services or the management of procedures? The answer is yes, with multiple examples already implemented across various regions of the country. In this article, we would like to explain three specific cases with small-scale pilots that have shown the potential impact of incorporating generative AI into public services. Let's dive in!

Generative AI Expert in Bureaucracy to Facilitate Procedures

In the first case, generative AI was used to improve knowledge management of internal procedures in the Sports Department of Alcorcón City Hall, Madrid. The project focused on centralizing and organizing information related to sports subsidies, a highly complex administrative process responsible for allocating 600,000 euros annually to various sports centers.

The solution was a chatbot that specialized in this type of grant and was able to guide each employee step-by-step through the stages of the process and their managers. The virtual assistant developed by the startup Dezzai, was trained on a wealth of relevant information, including legal documents, previous grants, and insights from civil servants. It uses natural language and an easy-to-use interface to answer users' questions.

"This tool went beyond what we initially expected, which was just a procedural manual. It raised an issue of internal reorganization," said Estela Fernández López, Head of the Technological Innovation Section at the City Hall.

‍

This AI specialized in grant bureaucracy functions as an internal knowledge repository that can be easily updated. It can also train new technicians in an agile, autonomous, and standardized manner. The pilot project also helped identify aspects that need to be considered if the intention is to extend the implementation of such solutions to other areas of the City Hall.

The project was part of GovtechLab Madrid, the innovation lab of the Community of Madrid, with which we've been collaborating for the past four years.

‍

Generative AI in Patient Case Monitoring

LOLA is a virtual assistant that autonomously makes phone calls to follow up with individuals who have visited the emergency room or patients undergoing chronic treatments in areas such as oncology and cardiology, among others. During these conversations, it collects information about their health status and transmits the data to the medical team. This technological solution was developed by Tucuvi, a startup that was part of the first edition of the Gobe Ventures accelerator.

What does an intelligent assistant like this mean for public administration? Patient follow-up is often a manual process, which consumes resources and time and can lead to potential delays. A generative AI automates the process, thus easing the workload of healthcare professionals, reducing unnecessary visits, and prioritizing cases based on their severity.

LOLA is used in more than 40 health systems across multiple regions and has an average patient satisfaction rating of 4.7 out of 5, according to the company's 2024 report. "It feels like you're talking to a person," said one.

‍

‍

In Spain, it has delivered positive results at El Baix Empordà Hospital in Girona, particularly in cases of chronic obstructive pulmonary disease (COPD). The outcomes include a 10% reduction in emergency visits, a 5.5% decrease in hospitalizations, and the complete elimination of readmissions within one month after discharge. "LOLA has revolutionized the care of COPD patients," stated the clinical team leader in the study's final report.

‍

Generative AI to accelerate Public Procurement Processes  

In the third example, we highlight a project by the Madrid City Hall that developed a generative AI solution to ease the administrative burden of creating public procurement documents.  

The tool, called BidGenius, was designed by WhiteBox, one of the winners of the Artificial Intelligence and Advanced Technologies Awards granted by the City Hall itself. It is an automated system that assists in generating public tender documents and simplifies the work of public officials in this process.

BidGenius uses a combination of procurement records, natural language models, and the retrieval-augmented generation (RAG) method. Based on the specifications entered by the public official, it analyzes previous tenders in the State Procurement Platform (PLACE) and selects the most relevant ones to automatically generate sections of the document, such as the Technical Specifications Sheet, while maintaining the formal tone and structure characteristic of public documents.

It has the potential to reduce the time it takes to create bid documents by up to 80%. This helps eliminate delays throughout the administrative process. It is currently being expanded to include other essential documents in the procurement process, such as the Administrative Clauses Document (ACD), a document outlining the terms and conditions of the contract.

Testing on a Small Scale to Fine-Tune and Evaluate Technological Solutions

In all three cases, a small-scale pilot methodology was used to evaluate the potential of AI and adjust its possible future scale. This approach aligns with many European initiatives aimed at adopting artificial intelligence systems, as highlighted in a study by the European Commission’s Joint Research Centre (JRC), led by Luca Tangi.

‍

These examples show how generative AI can transform public administration, particularly by improving time efficiency and internal processes. These targeted applications also help to enrich the debate on the risks and limitations of this emerging technology, providing important insights for its governance.

That’s why Gobe continues to support testing and adoption of these technologies in government. By piloting these solutions in public services, we aim to foster collaboration with startups and inform future technology needs. 

Do you face similar challenges where AI could make a difference? Are you leading a startup or digital SME with a product that has the potential to support public administration? Feel free to contact us or visit our platform, Open Gobe, to explore open calls for the entrepreneurial ecosystem.  

‍