Held at the Roche premises in Basel, Switzerland, on 15–17 September, the workshop covered use cases in generative artificial intelligence (GenAI) in GxP environments, predictive process modeling, digital twins, and precision medicines. Four highly interactive sessions were framed by keynote presentations, regulatory panel discussions and ample time for networking.
Over 130 participants had the opportunity to explore real, implemented digital use cases in pharma/biotech manufacturing and supply, and to discuss practical solutions to challenges in adopting digital technologies. Participants included an impressive lineup of regulators, both good manufacturing practices (GMP) inspectors and chemistry, manufacturing, and controls (CMC) assessors, from different European health authorities, namely the European Medicines Agency (EMA), Swissmedic, the Irish Health Products Regulatory Agency (HPRA), the Dutch Medicines Evaluation Board (MEB) and the Bavarian inspectorate.
Interactive Format Fostered Engagement and Collaboration
Inter¬active sessions were organized as world café style roundtable discussions. Participants were divided into four groups, identified by the colors of their lanyards, and were asked to gather at roundtables identified by numbers. Sessions started with short presentations of real world digital use cases on the four main themes of the workshop: GenAI in GxP environments, predictive process modeling, digital twins, and the use of digital tools in the manufacturing of precision medicines.
Following the flash presentations, roundtables in each group were asked to elaborate answers to several predefined questions focused on risk assessment/management, scalability, lifecycle management and inspection/submission relevant aspects. Table moderators entered consolidated answers into online forms which were further consolidated by a GenAI tool during a short networking break. The audience then reconvened to appreciate the consolidated output and address open points with use case owners and representatives of regulatory authorities in an interactive panel discussion.
Genuine thanks to the organizing committee for proposing this workshop format, which fostered a unique environment of open discussion, round tables and intensive exchange of knowledge. I do not recall attending a similar event where participants so readily set aside company affiliations and titles, engaging humbly with a genuine intention to contribute.
Alvaro Avivar-Valderas, Takeda
GenAI in GxP Environments
In the session on the use of GenAI in GxP environments, participants discussed how GenAI supports quality decisions following a use case presented by Laure Gurcel (Sanofi) and how GenAI can be leveraged to accelerate the deviation handling process following a presentation from Margo Banjedvorac and Marcin Chojecki (Roche). For the panel discussion, Sina Berndl (Swissmedic) and Roberto Conocchia (EMA) joined the stage. Participants agreed with the panelists’ assertion that AI assistants can be used in GxP environments for non-critical, supportive steps in quality business processes. Importantly, the panelists noted that a robust AI control strategy needs to be in place, and a “human in the loop” needs to verify outputs. They also noted that GenAI helps to shift human activities from mere data gathering and content generation to value added critical review, fostering critical thinking and emphasizing human accountability.
Predictive Process Models
In a session on predictive process modeling, use cases included feedback control methods in pharma manufacturing, by David Ochsenbein (Johnson & Johnson); smart yield optimization through self-learning AI engines, by Christos Michos and David Lauenstein (Takeda); and Critical Quality Attribute (CQA) multivariate analysis in cell culture, by Jonathan Ritscher and Arne Seeliger. Key takeaways from this session included the need to monitor data distribution and standardize data platforms to handle diverse inputs. Predictive models need to incorporate process input variability, and their performance should be predicated on continuous monitoring. Changes to models should be documented and lifecycle management tools validated.
Digital Twins
In the session on digital twins, participants discussed AI-powered end-to-end digital twins for adaptive process control, presented by Thomas Zahel (Körber Pharma), and predictive control of cell culture experiments in automated bioreactors, presented by Jean-Marc Bielser (Merck). Marcel Hoefnagel (MEB and chair of the EMA Quality Innovation Group), joined the panel discussion. Here again, data preparation and solid data governance were highlighted as important elements of model development and deployment. Participants agreed that digital twins hold significant potential for simulating and optimizing complex manufacturing processes, supply chains, and facility designs to save costs and prevent failures.
Digital Tools for Precision Medicines
In the last session on the use of digital tools for precision medicines, use cases involved using AI in advanced therapy medicinal products (ATMP) manufacturing, presented by Alvaro Avivar-Valderas (Takeda); statistical process control and control limit calculations in cell therapy manufacturing, presented by Manuel Torres (Takeda); digital enablers for decentralized manufacturing of cell therapies, presented by Renato Panza (Galapagos); and using machine learning to make individualized mRNA immunotherapies, presented by Don Low (Genentech/Roche). Hoefnagel also joined the panel discussion. Roundtable discussions underscored the importance of a robust data governance framework for individualized therapies. Presenters noted that standardizing data and processes is fundamentally important, and the focus should be on data quality, integrity, data privacy, and a reliable data flow. For individualized therapies, the presenters asserted that scale can be achieved through centralized or decentralized manufacturing models, the choice of which will depend on modalities.
Conclusion
The workshop was also a huge success thanks to Roche’s generous and warm hospitality. Participants had the opportunity to view latest developments in digitalization during a gallery walk and virtual site visits. Evening networking events on top of the Roche towers and on board a Rhine River cruise ship nicely complemented the highly interactive three days.
This year’s ISPE D/A/CH Affiliate workshop underscores ISPE’s role as a catalyst for advancing digital maturity across the pharma industry. As the digital journey continues, events like this enable all stakeholder to weigh in and provide the resources needed to navigate complexity and drive innovation.
The planning committee members hope that the outcome of workshop will positively impact regulatory guidance that is currently in development or under review, namely EU GMP Annexes 11 and 22. The team plans to publish a more detailed analysis of workshop outcomes in the ISPE Pharmaceutical Engineering® magazine in early 2026.
Acknowledgements
ISPE D/A/CH workshop Planning Committee: Viktor Mettler, Takeda and ISPE D/A/CH Board member (Chair); Ursula Busse (Co-Chair); Laura Kuger, Takeda; Roger Nosal, Vaxcyte; Nick Lee, HPRA; Andreas Kleeb, Roche; Christian Wölbeling, Körber Pharma and ISPE D/A/CH Board member; Arne Seeliger, Roche; Margaux Penwarden, Roche
