Sue Tabbitt, senior writer at Sarum Life Sciences reflects on a tumultuous year for technology strategy delivery in life sciences.
2025 has been a trying year in life sciences. “Both geopolitics and the promise of AI have influenced the dynamics in this industry,” notes Frits Stulp, partner for life sciences at Implement Consulting. “We’ve seen an increase in reorganisations and lay-offs, but at the same time a renewed drive to deliver process improvements in the development and approval of new medicines.”
AI’s Deepening Promise
The challenging market has helped to highlight the scope of advanced technologies, certainly. “The power of AI is undeniable,” says Remco Munnik, owner and founder of Arcana Consultancy, highlighting the potential to change the way regulatory teams work. “What once required custom-built data mining tools (e.g. robotic extraction of SmPC and Module 3 sections) is now standard functionality in modern AI platforms,” he notes. “This shift from bespoke tooling to scalable intelligence is a game-changer for regulatory operations.”
But to fully harness the possibilities, operational teams, systems and data need to be properly prepared. That means good integration; complete, correct, standardised data; and professionals who know how to use the technology appropriately.
“Where companies have struggled to produce ‘immediate’ AI benefits, they have found that having a clear data strategy with strong data governance practices cross-functionally is paramount to maximising AI potential,” notes Steve Gens, founder of Gens & Associates, which tracks the advancement of life sciences’ product data management via periodic global surveys.
Next-Level Dossier Automation
Where leading companies have put the data work in to enrich their regulatory data, Gens is observing a shift towards dossier module automation as the next target for process improvement. Here, companies are looking beyond individual document generation to multi-purpose automation across investigational, new marketing, variation and other use cases. “The focus this year has been on clinical, safety, and CMC document generation, along with health authority query ingestion automation,” he says.
Combining document generation, document quality control, and translation automation constitutes the next-level of regulatory information management transformation, Gens says – with an implementation tipping point expected in 2026/2027.
The Rise of Agentic AI
For Jason Bryant, senior VP for AI-based product development at ArisGlobal, companies that are serious about harnessing AI to enhance their impact and value are seeing the rewards now. “Already, AI is enabling new cost-efficiency in regulatory affairs and drug safety/pharmacovigilance,” he says. “Positive use cases to date include marketing authorisation application preparation, product change control/regulatory impact assessment management, adverse event case processing, and safety reporting.”
But there is even more to play for with agentic AI, he notes. This has come to the fore as a desirable strategic capability in 2025, because of the multiplied benefits of employing co-ordinated AI ‘agents’, capable of reasoning to achieve defined goals in the most optimal way – and even suggest different ways of working, or alternative objectives and ways that teams, and companies, can add value.
In a regulatory context, opportunities for agentic AI include reinventing the global management of product regulatory compliance, Bryant says. “Autonomous, regulation-aware dossier assembly and submission orchestration is within reach. It’s possible for orchestrated AI agents to continuously ingest clinical data packages, study reports, CMC documents, eTMF pointers and legacy submission artefacts. Agentic systems could also perform automated regulatory gap-analysis versus target-region requirements, draft region-specific CTD/eCTD modules (with citations and traceability to source documents), and orchestrate the technical packaging (file naming, folder structure, etc).”
But again, this relies on companies having put in the requisite groundwork. Beena Wood, a drug safety specialist and chief product officer at Qinecsa Solutions, says 2025 has seen a reality check in this regard. In a pharmacovigilance context, she notes. “The impressive organisations have stopped chasing shiny AI pilots and started investing in ‘boring’ foundations: data quality; standardisation; and governance,” she says. “They’re not deploying fancy agentic systems yet – but they’re positioned to succeed when they do.”
Cultivating AI Confidence
Through their early deployments of AI, companies have learnt a lot about what’s possible – and appropriate. This is allowing them to be bolder and more intentional in their application of the technology – in whatever form they are ready to deploy it.
ArisGlobal’s Bryant emphasises that including human expertise will remain crucial, but will increasingly involve routing suggestions for human review when potentially ambiguous scenarios arise and an expert check is needed.
In emerging deployments, governance should serve as a facilitator of progress as well as a mitigator of risk, if companies are to see step changes in productivity and better outcomes. Taking a principles-based approach, rather than one that is hard-wired around specifics, can better support process stakeholders in defining scenarios and goals that agentic AI can help solve, and how human-on-the-loop provisions might best evolve.
Keeping Patient Outcomes Front and Centre to Evolving Ambitions
So what lessons can companies take forward, as a new year beckons? For Implement’s Stulp, the potential now is to rethink the substance of the work being done, and about the outcomes of any transformation – e.g. for patients. Implement Consulting is tracking developments such as those around cloud-based “Trusted Regulatory Spaces”, designed to accelerate global drug approvals by fostering more dynamic and streamlined collaboration between pharma companies and regulators, as advocated through Project PRISM[1], an initiative born out of (but not exclusive to) the FDA. Says Stulp, “The scope here to accelerate process and system improvements is considerable, and goes hand-in-hand with the promises of process optimisation through AI.”
In the meantime, focusing on culture and strategy as much as technology’s growing sophistication is imperative. Qinecsa’s Wood says one of her biggest disappointments in 2025 has been insufficient investment in people and digital mindset transformation, something that applies across most industries. “Billions have been poured into AI technology yet companies have underfunded the human capability development required to make it work,” she says. “The balance needs to be redressed in 2026.”
[1] Cloud Transformation through Trusted Regulatory Spaces (TRS): https://hello.dnanexus.com/webinar/cloud-transformation-through-trusted-regulatory-spaces
