The life sciences industry has spent the better part of a decade modernising its regulatory infrastructure — investing in global regulatory information management (RIM) systems, improving data quality and, more recently, committing significant budget to AI. Yet the findings of a new Operational Excellence and World Class RIM℠ study, covering 59 pharmaceutical, biologics, medtech and gene therapy organisations, suggest that most of that investment could be built on shaky ground.
The study, now in its 47th cycle, is the life sciences industry’s longest-running regulatory benchmarking programme. This time, it introduced a Future Readiness Indicator (FRI) alongside its gauges of established performance. This uncovered a sobering finding that only one of the participating organisations qualified as ‘ready and leading’ for an impending period of significant change. More than a fifth (21%) were deemed ‘at risk’ of not being sufficiently prepared, while the remaining 77% were making progress but still have clear gaps to address.
The Data Problem Underlying AI Ambitions
Of all the findings in this year’s study, the data accountability results are the most striking. Only four of the 59 organisations surveyed has by now genuinely embedded a culture of holding individuals and teams explicitly accountable for the accuracy and quality of data held in regulatory systems. This has had a significant bearing on their trust in that data and what they are able to do with it.
These organisations recorded an aggregate data quality confidence score of 93%, for instance, against 50% for the remaining 55. Their efficiency across 15 core RIM capabilities was 93%, versus 70% for the rest. For health authority commitment tracking data specifically, the situation was even starker: 100% high confidence vs 44% of the majority.
For organisations investing in AI, this is significant. AI tools applied to poorly governed, inconsistently owned regulatory data will not produce reliable outputs — and in a regulated environment, unreliable outputs are not a minor inconvenience. This needs to be addressed, through tighter accountability for data. Historically, regulatory functions have organised ownership around documents and dossiers, where authorship and approval chains are well established. Extending that accountability to individual data elements — who is responsible for the accuracy of a specific field in a regulatory system — is newer territory that now needs to be covered. Almost two-thirds of organisations in the study report actively working towards this goal, but few have achieved it so far.
A Reality Check on AI Adoption
47% of companies in the study say they have AI pilots or implementations underway, and every large company claims to be making significant AI investments. Yet, of 131 benefit-realisation responses tracked across all AI and advanced automation use cases, just six exceeded expectations. Broad consensus on realistic implementation timelines has shifted to 2027–2028 as a result.
Where AI is delivering results today, the conditions enabling this consistently involve well-defined processes, structured and trusted data and a clearly delineated task scope. Clinical document generation is the leading use case, followed by AI-assisted translation and CMC content authoring. Although the potential to compress the timeline from clinical study closure to regulatory filing is real, it is contingent on the data governance and process infrastructure being in place before the AI capability is deployed (it cannot be built alongside it).
A broader shift — AI-assisted generation of full dossier sections rather than individual documents — is projected to become viable at scale around 2028. The organisations best placed to reach that inflection point will be those that have already done the less visible work: governing their data properly, maturing their processes and embedding accountability at the right organisational levels.
Process Maturity as a Technology Readiness Issue
This latest research cycle introduced formal process maturity measurement for the first time, using the Capability Maturity Model Integration (CMMI) framework across nine core regulatory process areas. The findings frame process discipline not as an operational nicety but as a direct driver of technology readiness.
Top performers predominantly operate at Level 4 (measuring) or Level 5 (optimising) for most processes. The broader group sits at Level 3 or below. The performance gap is substantial: top performers report operational throughput improvements at 80% versus 47% for peers; operating cost improvement at 90% vs 39%; user productivity at 90% vs 50%.
Organisations operating through KPI-driven continuous improvement have the structural flexibility to absorb new capabilities as they mature. Those still at Level 2 or 3 are carrying process debt into a period of accelerating change — and each new technology wave, whether related to AI, structured data mandates or cloud-based regulatory spaces, will expose that debt further.
Cloud-Based Regulatory Spaces: a Structural Shift Coming Fast
One development from this year’s study warrants particular attention from a health technology perspective. Cloud-based regulatory spaces (CBRS) — shared digital environments in which regulatory data is exchanged between industry and health authorities — have moved from concept to active adoption far faster than most anticipated. 41% of organisations are already participating in a CBRS initiative; a further 47% plan to do so within a year; and 71% believe it will fundamentally change how the industry works with health authorities within five years.
Early pilots have demonstrated technical feasibility for a model in which a single dossier is reviewed simultaneously by multiple regulators, replacing today’s sequential, one-to-one submission process.
The data infrastructure implications are significant. CBRS depends on clean, structured and consistently governed regulatory data. Organisations with weak data foundations are likely to find that participating effectively in these environments is harder than the technology itself suggests.
The Gap Between Investment and Readiness
Across every dimension measured in the 2025 study — data quality, process maturity, AI adoption and structural readiness for CBRS — a common pattern exists. Organisations that have invested in governance, accountability and process discipline are pulling away from those that have focused primarily on technology. Although technology investment is necessary, by itself it is not sufficient.
For a sector now committing seriously to AI, that finding has practical implications. The organisations that will extract the most value from AI in regulatory work over the next 3-5 years won’t necessarily be those spending the most on the associated tech today. They will be those that have built the requisite data infrastructure and organisational discipline for AI to be able to function reliably in a regulated environment.
About the author
Steve Gens is Managing Partner of Gens & Associates Inc, a life sciences benchmarking and advisory firm specialising in regulatory information management. The full 2025 Operational Excellence and World Class RIM℠ study white paper is available here https://gens-associates.com/2026/04/03/2025-operational-excellence-and-world-class-rim-study-whitepaper/ .