Event presentation by: Marco Orellana, CIDAI Manager
11:30 | Institutional Welcome
- Joan Mas i Albaigès, Director of CIDAI and Scientific Director of Digital Technologies at Eurecat
- Gemma Roig, Director of Organisation, Innovation and Transformation, Grup Caixa d’Enginyers
- Arnau Serra, Head of the Technical Service Centre of the Secretariat for Digital Policies, Government of Catalonia
11:45 | Innovation in Experimentation and Detection of Financial Fraud: How AI Strengthens Trust in Financial Transactions
Marta Regina Cano, Director of Data and Transformation, Grup Caixa d’Enginyers
She explained how the project reached the organisation and how it was developed. She presented the strategic vision of the project from the perspective of data‑driven innovation, collaboration between research and industry, and the value that applied artificial intelligence brings to fraud prevention. She highlighted a culture based on data and continuous improvement in member protection, and how it returns tangible value to the organisation —both in knowledge and in analytical maturity and capacity for collaboration with the scientific ecosystem.
Rosa Olivares, Specialist at the Anti‑Fraud Unit, Grup Caixa d’Enginyers
She contributed the operational and security perspective, explaining how the research results will be integrated into current fraud detection and response mechanisms. She detailed the practical benefits of the new system, its real‑time alerting capabilities, and how it enables more proactive and efficient action in the face of potential threats or anomalous behaviour, complementing traditional prevention systems. Her intervention highlighted the importance of combining expert knowledge, innovation, and technology to advance towards more proactive, scalable, and human‑centred fraud management.
11:55 | Fraud Prediction System: Anomalous Patterns in Economic Transactions
Àlex Casamitjana Serra, Researcher at the Big Data and Data Science Unit, Eurecat
To introduce the project from a technical perspective, he began with a brief review of the available data sources, the initial explorations conducted, and their importance in the modelling process. With these elements, he explained the chosen algorithmic approach to address the fraud prediction problem. He concluded by detailing the tasks associated with dataset preprocessing, an essential part of the project.
Cirus Iniesta, Director of the Big Data and Data Science Unit, Eurecat
This presentation detailed the modelling process and the evaluation methodology used. He then presented the results obtained throughout the different phases of the project. Special mention was given to the microservice developed to bring the use case closer to production, making it scalable and reusable beyond the current project. He closed with a series of reflections and lessons learned throughout the project.
12:25 | Member Profiling and Anomaly Detection at Grup Caixa d’Enginyers
Karina Gibert, Professor at UPC‑BarcelonaTech and Dean of the Catalan Professional Association of Computer Engineering (COEINF)
From IDEAI, member profiling was analysed based on data provided by Grup Caixa d’Enginyers. Seven differentiated typologies were identified according to the types of products contracted and the duration of the relationship with the institution. These profiles were obtained using a proprietary methodology combining innovative clustering techniques with advanced profile‑interpretation procedures.
The proof of concept was carried out with a specific sample of salaried members from the province of Barcelona. From this sample, transaction patterns associated with the different profiles were analysed, and anomaly‑detection mechanisms were designed to identify potentially suspicious operations not yet covered by the bank’s current business rules.
12:40 | Open Q&A
13:00 | End of the Event
