For more than a decade life sciences organisations have spoken confidently about becoming data-driven; more recently about becoming AI-powered. Yet compared with other highly-regulated sectors, the pharma industry still struggles to exploit its data in a strategic, joined-up way, which could now impede their scope to fully exploit AI.
A recent roundtable discussion brought together experts in the field to ask why progress has been so slow, and what needs to change now. On the debate panel were Remco Munnik, owner and founder of Arcana Life Sciences Consulting; Frits Stulp, partner for life sciences at Implement Consulting Group; and Peter Brandstetter a senior manager at Accenture. ArisGlobal’s Ian Crone, who chaired the event, reports on the key points raised.
A recurring theme throughout the discussion was whether pharmaceutical companies are genuinely addressing the root cause of their data challenges, or merely responding tactically to regulatory pressure.
A decade of missed opportunity
The ISO Identification of Medicinal Products (IDMP) standards were intended to be transformative, defining how medicinal products should be uniquely identified and described using structured, standardised data. Conformance with these standards should enable seamless data exchange between companies, regulators and healthcare systems, while supporting improved patient safety and operational efficiency.
In the EU, IDMP has been on the regulatory horizon since about 2012, yet repeated delays, evolving specifications and uncertainty about enforcement timelines have eroded industry momentum over the last decade or more. Rather than using the time to modernise their data foundations, many organisations chose to wait. That decision is now proving costly. As Peter Brandstetter observed during the discussion, speaking for the industry as a whole, “We are already too late. We should have started 10 or 15 years ago.”
Other industries, he argued, have used comparable timeframes to build integrated product lifecycle management capabilities and unified data models that serve far more than compliance needs. Life sciences, by contrast, has typically delivered only what was required “to pass the next deadline”.
This short-termism has left many companies with fragmented systems, duplicated data and unresolved debates over which system represents the authoritative version of product information.
In Frits Stulp’s view, these issues have not only driven cost overruns and failed programmes, but in some cases have had personal consequences. He pointed to careers derailed by rushed implementations and immature requirements – projects that addressed yesterday’s regulatory needs while creating tomorrow’s structural problems.
AI enthusiasm meets fragile foundations
Despite unresolved data challenges, enthusiasm for artificial intelligence is accelerating rapidly across the sector. Today, AI is being explored for activities ranging from safety signal detection and submission authoring to labelling harmonisation and portfolio analysis. Early pilots have delivered encouraging results, adding to the sense that AI could reshape regulatory and operational work.
However, the panel repeatedly stressed that AI cannot compensate for weak data foundations. There is a widespread belief that large language models or generative AI can infer structure, correct inconsistencies or “fill in the gaps”. In a regulated environment, that assumption is dangerous.
“AI really requires good data,” Remco Munnik noted. “Without structure – without governance that provides meaning – AI struggles to make sense of information.”
Brandstetter echoed this concern, warning that the current AI hype cycle is encouraging organisations to experiment before they are ready. Without consistent, trusted data, he said, AI initiatives risk producing unreliable or misleading outputs—ultimately damaging confidence in the technology itself.
The message is clear: the more effort companies invest in building high-quality, standardised and well-governed data, the more value AI can unlock. There are no shortcuts.
Reframing regulatory affairs as a data asset
Considering a constructive way forward, the panel homed in on the evolving role of the regulatory affairs function. Traditionally viewed as document-heavy and compliance-driven, regulatory teams are rarely positioned as strategic contributors to enterprise data initiatives. Yet they sit on some of the richest, regulator-validated product information within any pharmaceutical organisation, the panel noted.
Stulp argued that, with the right mindset and structure, regulatory affairs could become a genuine “data powerhouse”. Clean, standardised regulatory data has the potential to do far more than support submissions. It could underpin portfolio planning, lifecycle strategy, market intelligence and even intellectual property decisions.
AI should be seen as an enabler, not a replacement, for regulatory professionals, Munnik suggested. He described a prototype scenario in which structured product data enabled automated downstream propagation of approved Company Core Safety Information (CCSI) changes. Updates flowed consistently through English and local labelling, patient leaflets and multiple translations. The result: improved efficiency and consistency, alongside the elimination of expensive, repetitive manual translation work. Rather than reducing headcount, such an approach would elevate the role of specialists, freeing them to focus on judgement-based and strategic tasks.
IDMP as an enabler, not an obligation
From a regulatory perspective, the European Medicines Agency’s Product Management Service (PMS) emerges as a critical piece of infrastructure. The panel described PMS as the linchpin of Europe’s transition towards structured regulatory data – supporting use cases such as shortage management, electronic application forms and the planned replacement of XEVMPD.
According to Stulp, PMS has reached a level of maturity that shifts responsibility firmly onto Marketing Authorisation Holders (MAHs). As Munnik put it, EMA has “done its homework”. The onus is now on companies to align, enrich and validate their data.
Seen more broadly, IDMP provides a common language for product data – not just for European submissions, but across internal functions and global markets. For AI, this consistency is essential. For regulators, it will enable more efficient oversight. For patients, it promises faster access to clearer, more reliable information.
Causes of inertia, and how to regain ground
Reflecting on why so many organisations have struggled to realise these benefits, the panel identified several recurring patterns ranging from fragmented leadership and programme “drift”, to technology-first thinking, and minimum compliance approaches to data transformation.
The panel agreed that it is not too late to turns things around, however. Companies could get back on track by validating their data against EMA PMS, resolving mismatches in product families, preparing for manufacturing and packaging enrichment, and working closely with vendors to ensure transparency and interoperability (e.g. as they move towards more fluid data exchange and more systematic adoption of AI).
Equally important, the panelists noted, is the need for a long-term data vision that extends beyond regulatory affairs and is delivered through incremental, value-led steps. “If companies begin this work now, they could still catch up,” Brandstetter suggested.
Looking ahead, Stulp highlighted the emergence of “trusted regulatory spaces” – shared digital environments where regulators and industry collaborate on data and review processes. Brandstetter went further, envisaging a future in which fully data-driven submissions could lead to near-instant approvals.
For patients, Crone concluded, the prize of all of these efforts should be better-quality information, delivered faster and in more accessible ways – an outcome that will matter even more as AI infiltrates every aspect of people’s lives.
By Ian Crone is VP Europe & APAC Regulatory Solutions at ArisGlobal.
The panel discussion, which took place in late 2025, was the latest episode of the Life Science AI Exchange Podcast, sparked by ArisGlobal. www.arisglobal.com
