Introduction

In response to growing competitive pressures — and amid a broader wave of digital transformation across the biopharmaceutical sector — R&D leaders are increasing investments to boost the efficiency, productivity and quality of their regulatory processes. Generative AI (GenAI) is emerging as a key part of this shift, offering new ways to enhance regulatory performance. However, while enthusiasm for AI is high, realizing tangible value remains a challenge. A recent BCG study found that 75% of executives ranked AI and GenAI as a top-three strategic priority, yet only 25% reported seeing significant value from these technologies1. This gap underscores the need for organizations to move beyond experimentation and invest in structured, end-to-end transformation to unlock GenAI’s full potential.

To better understand how GenAI is being explored within regulatory functions, BCG convened a roundtable (the 2024 Consortium for Regulatory Affairs Leaders (CORAL) roundtable), comprising 28 senior executives from 20 leading biopharmaceutical companies. Insights from the roundtable discussion and a supporting benchmarking survey (completed by 17 companies) inform key themes covered in this article, including high-potential use cases, expected benefits and the key enablers for success.

GenAI gaining momentum in regulatory functions

The benchmarking survey, described in detail in the Supplementary information, revealed that all 17 survey participants are actively experimenting with GenAI tools (up from just half in November 2023). Notably, 12% have begun rolling out initial solutions, including use cases such as synthesis of regulatory intelligence and drafting of labelling documents. This compares to none in November 2023.

Although this finding illustrates GenAI’s accelerating momentum, most biopharma companies are still in the early stages of adoption, with ~85% building minimum viable products or proofs of concept (Supplementary Fig. 1). Additionally, 70% of participants reported implementing GenAI tools on a use case-by-use case basis, which is another indication of lower maturity. A handful of larger biopharma companies are now shifting to a more comprehensive, end-to-end GenAI-enabled approach.

GenAI regulatory use cases and benefits

When asked about the top-priority GenAI use cases, participants cited regulatory intelligence management (71%) and document authoring (59%) (Supplementary Fig. 2). GenAI can help generate regulatory intelligence by analysing large volumes of regulatory guidance, historical submissions and health authority feedback to identify patterns in agency preferences and factors associated with successful outcomes. These insights can inform regulatory strategy, including the choice of submission pathway and the timing and sequencing of submissions, while also helping teams anticipate likely agency queries.

Beyond regulatory intelligence, GenAI has shown strong potential in automating document authoring, dossier preparation and clinical trial documentation. For example, it can generate high-quality first drafts of electronic common technical document (eCTD) modules, informed consent forms and patient information leaflets. One top 10 pharma company shared during the roundtable that its use of GenAI reduced writing times by 50%–70% for study reports, protocols and labelling documents.

Roundtable participants noted that GenAI has the potential to enhance not only the speed, but also the quality and consistency of outputs. When guided by regulatory context, such as health authority guidance, it can help ensure alignment with agency expectations, and enforces consistent structure and formatting, independent of the author.

Some companies are also exploring GenAI tools in regulatory-adjacent areas such as advertising and promotional material review. These traditionally complex, resource-intensive workflows can be substantially enhanced through GenAI integration. GenAI can help analyse clinical data and references to evaluate the validity of proposed claims, supporting the development of structured claims libraries. GenAI can also recommend content refinements and revise draft materials — freeing up regulatory teams to focus on higher-value, strategic activities.

Overall, GenAI has the potential to enable more efficient decision-making, improve quality and accelerate submissions — while helping to reduce the risk of non-compliance due to manual errors, content inconsistencies and misalignment with agency expectations. Survey participants reported that they expect the implementation of GenAI solutions will result in substantial improvements in speed (in some cases, up to 50% acceleration) and productivity (61% expect significant or very significant gains) (Supplementary Fig. 3). Speed, productivity and quality were the primary motivators, whereas direct cost reduction was not a central focus for most participants.

Common pitfalls and approaches to addressing them

The integration of GenAI into biopharma regulatory functions such as document authoring, submission planning and regulatory intelligence offers transformative potential, but it also introduces key challenges that must be carefully managed.

A key risk is that GenAI may produce inaccurate or misleading content, known as ‘hallucinations’, particularly when outputs are not grounded in regulatory context. These hallucinations often appear plausible, making them harder to detect, and can compromise submission integrity — that is, the accuracy, consistency and compliance of materials submitted to health authorities. Mitigation strategies include strong data governance, domain-specific fine-tuning and oversight by regulatory subject matter experts — not only to review outputs, but to frame the right questions, structure prompts appropriately, and interpret results within a regulatory and scientific context. Capabilities such as citation mapping, version control and content traceability, tracking both the source and evolution of generated content, can further support auditability and build trust in AI-generated outputs.

Fine-tuning GenAI models on internal company data is important across all regulatory use cases, given differences in terminology and content structures between organizations. This is particularly critical for chemistry, manufacturing and controls (CMC) content, where publicly available data are limited and often heavily redacted. As a result, applying GenAI to CMC authoring will likely require a higher degree of reliance on internal data and tailored model development than other areas.

Regulatory professionals may still be hesitant to rely on AI-generated content, fearing a loss of control or increased risk of regulatory non-compliance or inaccurate submissions. This hesitation can be addressed by positioning GenAI as an assistive tool that augments, rather than replaces, human expertise. Providing comprehensive training, involving end users in piloting use cases, and highlighting early wins can also help build trust and confidence. When these elements are in place, organizations are better positioned to deploy GenAI safely and effectively, thereby accelerating innovation, while safeguarding compliance.

Critical factors for successful adoption of GenAI

Roundtable attendees highlighted three critical success factors for the implementation and adoption of GenAI solutions: leadership focus, a robust data foundation and a robust end-to-end approach. But successful transformation goes beyond technology. Top-performing organizations follow the 10–20–70 principle, dedicating 10% of their efforts to algorithms, 20% to data and technology, and 70% to people, processes, cultural transformation and change management1.

94% of participants highlighted leadership vision, focus and prioritization of GenAI as a critical success factor; 71% cited a strong business case. Participants discussed the importance of building a robust data foundation, with emphasis on data governance and master data management to ensure accuracy, accessibility and reliability.

To capture GenAI’s full potential, companies must shift away from single use cases and a series of isolated solutions. Instead, organizations should adopt an end-to-end approach that is informed by their key business goals. Transformative end-to-end solutions offer benefits such as generalizable templates, flexibility for non-technical users and automatic updates of dependent documents. To realize these benefits, the different technologies need to complement and build upon one another (such as traditional enterprise software coupled with AI/ML and GenAI).

The key to a successful end-to-end GenAI implementation is change management, with 76% of participants citing it as the most important enabling capability, followed by GenAI technological literacy (65%) and a strong IT–business partnership (65%) (Supplementary Fig. 4). Leaders across the organization must communicate the benefits of the change, identify and support early adopter champions, and offer guidance and resources to employees. Equally important is the establishment of an iterative process that systematically measures performance and solicits feedback from end users for continuous quality improvement and sustained use of GenAI at scale.

Conclusion

This is a crucial time for biopharma regulatory teams to invest in GenAI, enabling them to do more with the same resources and, consequently, focus on being a strategic partner to R&D and the business. Companies that do not seize the GenAI opportunity will inevitably fall behind.