On 15 and 16 September 2026, the UIC Opt-in Project “Artificial Intelligence Models, Scalability and Transferability” (AIST) brought together railway organisations in Inchicore, Dublin, for a workshop hosted by Irish Rail (IE).

The workshop brought together representatives from Irish Rail, German Railways (DB), Austrian Federal Railways (ÖBB), the Norwegian National Rail Administration (Bane NOR), Railway Infrastructure of Portugal (IP), the Italian Railway Network (RFI) and UIC, with the Spanish Railway Infrastructure Manager (ADIF) joining remotely. The two-day programme focused on practical experiences and the conditions needed to scale AI projects across railway organisations.

Among the topics discussed was an AI-based predictive maintenance project for points machines, using a third-party solution, as well as approaches to structuring asset data based on reliability centred maintenance (RCM) principles to support fault analysis and trend identification in infrastructure maintenance systems.

Participants also examined the main factors blocking or facilitating the scaling of AI projects, including the use of stage gates, and discussed a framework for evaluating an organisation’s AI readiness and maturity.
The programme also included practical demonstrations and site visits, providing an opportunity to link technical discussions to operational applications. Participants attended a train simulator demonstration, visited the assembly site for new electric and battery trains, and took part in a site visit and demonstration of a new signalling control system.

The workshop also provided an opportunity for railway organisations to share experiences, identify common challenges and explore approaches that can support the wider scalability, explainability, and transferability of AI solutions across the sector.

More information on the AIST project is available on the UIC website.