AI initiatives often get stuck in the proof-of-concept stage and fail to scale.

In this session, we will see real-world examples of how companies have managed to bring artificial intelligence solutions to their production environment in different sectors, generating operational or strategic impacts.

This session will help us to understand the key success factors, common barriers and lessons learned on the path from innovation to implementation.

Presenter:

• Carolina Migliorelli, Head of Research Line (Healthcare Artificial Intelligence), Eurecat

Participate:

• Pau López, Head of Information Systems, Catalan Health Institute (ICS)

• Carles Rubies, Director of Digital Transformation, Granollers General Hospital

• Oriol Canillas, IT Technician & Ernest Ceballos, Data Analyst, SOME S.A.

• Pol Colomer, Co-Founder & Co-CEO, Dribia Data Research

eurecat.org | @Eurecat_news | Instagram

Dr. Carolina Migliorelli Falcone is the head of the research line on Trustworthy Artificial Intelligence for Healthcare within the Digital Health Unit at Eurecat. Her research career began with a doctoral thesis at the Center for Biomedical Engineering Research (CREB – UPC), followed by a postdoctoral fellowship at the Network Center for Biomedical Research (CIBER-BBN).

Her work focuses on the development of artificial intelligence systems for healthcare, with a particular emphasis on trust, explainability, and security. She leads projects aimed at creating advanced machine learning algorithms that support clinical decision-making, facilitate patient classification and stratification, and promote data-driven, personalized interventions. She also works on solutions that empower individuals to manage their health and improve their lifestyle through reliable and user-friendly digital technologies.

She has extensive experience in biomedical data processing and analysis, as well as in integrating clinical data in complex environments. She holds a PhD in Biomedical Engineering (UPC), a Master’s degree in Biomedical Engineering (UB-UPC), and a Bachelor’s degree in Telecommunications Engineering with a specialization in Electronic Systems (UPC).

Institut Català de la Salut (ICS) | Linkedin

Pau López graduated as a Telecommunications Engineer in 1999. Since then, he has held various roles in healthcare information systems across several institutions. In 2019, he led the creation and deployment of the DigiPatICS project, focused on digitizing the Pathology Services of the Catalan Health Institute (ICS). This initiative also paved the way for the development of AI algorithms applied to cellular imaging, positioning the project as a European leader in digital pathology. He currently serves as Head of Information Systems at ICS and supports corporate initiatives related to the implementation of artificial intelligence solutions.

The digitization of pathology slides within the DigiPatICS project has opened new opportunities for the development of artificial intelligence algorithms, aimed at cellular quantification and tumor pattern recognition, among other applications. Several of these algorithms are already integrated into the daily workflow of pathologists. They have been developed in collaboration with the Polytechnic University of Catalonia (UPC), with the support of pathology professionals, using various AI approaches to ensure clinical applicability.

Hospital General de Granollers | Linkedin

Carles Rubies is a mathematician and expert in information technologies applied to health, with more than 30 years of experience in digital healthcare innovation. He has led pioneering projects in medical imaging systems and, currently, as CIO of the General Hospital of Granollers, he drives the digital transformation and implementation of Generative Artificial Intelligence in an institution with more than 2,000 professionals. He has deployed clinical digitalization projects, BI, cybersecurity and Artificial Intelligence initiatives, integrating innovation with real impact on care and operational efficiency.

From testing to real deployment, with impact. The General Hospital of Granollers is implementing the use of Generative AI in real clinical environments. AI agents are being integrated into day-to-day operations to write discharge reports, prepare Clinical History summaries and respond to complaints, among other things. With leadership from digital transformation, co-creation with professionals and an active culture of training and continuous improvement, the project has generated a high operational and cultural impact, demonstrating that it is possible to scale AI with meaning, efficiency, quality and sustainability.

www.some.es | LinkedIn

www.some.es | LinkedIn

Computer Systems and Networks Technician, currently IT Intern at SOME S.A.U. He has participated in projects involving the integration of touch screens, RPA, and Machine Learning to improve production processes and reduce non-conformities.

Graduated in Computer Engineering from the University of Girona (UdG) and currently IT Intern at SOME S.A.U. Passionate about data analytics, he works on Machine Learning projects applied to industry. Starting in 2025, he will begin a Master’s in Bioinformatics at Aarhus University.

In this talk, we present an AI project developed at SOME S.A.U. to predict defects in metal parts in real time. Using data collected during production and models such as LSTM and HMM, promising results have been achieved to improve quality, reduce costs, and scale the solution within the framework of Industry 4.0.

www.dribia.com | LinkedIn

Pol Colomer holds a PhD in Physics focused on complex networks from the University of Barcelona, with research experience at UC San Diego. After academia, he spent over six years as a data scientist at King—the company behind Candy Crush—optimizing mobile games through analytics and machine learning. For the past eight years, he has been co-founder and co-CEO of Dribia, a Barcelona-based data science studio that designs and delivers tailored AI solutions for companies across industries.

The Daima project emerged from Celsa’s need, a leading steel manufacturer, to automatically classify the quality of scrap metal received. This algorithmic solution, based on artificial intelligence and deep learning, utilizes convolutional neural networks to analyze images from static cameras and identify different material qualities in real time. The system, which emulates manual work, improves operator safety, increases the objectivity of the assessment, and streamlines the process, with the added benefit of prediction explainability.