Potential to performance: How regulatory organizations are adopting AI
This article reports the findings from a 2025 benchmarking survey on the adoption of regulatory artificial intelligence (regulatory AI) across pharmaceutical and life sciences industries to learn how regulatory affairs organizations are turning their AI ambitions into action. Industry leaders are investing in AI solutions for regulatory intelligence, drafting assistance, and research assistance, unlocking measurable gains in efficiency, compliance, and insight. Although adoption remains uneven, with many organizations seeking clearer evidence of return on investment (RoI), some are realizing measurable impact. The findings highlight the emerging practices and strategies defining the next era of regulatory excellence and how organizations can improve implementation and scale readiness during the AI revolution.Keywords – agentic AI, AI investment, bioconvergence, generative AI
Introduction
AI is reshaping the life sciences industry faster than most organizations can adapt.1,2 Recent studies show that life sciences organizations are adopting AI across the research and development continuum ranging from clinical trial design and execution to pharmacovigilance and regulatory documentation. However, most deployments remain in early or pilot stages, emphasizing applications in automation, data analysis, and decision support rather than fully integrated enterprise use.1-3 In regulatory affairs, the pressure is mounting with ever-expanding data volumes and submission requirements intensifying, driven by the expectation of shortened innovation cycles. However, as more companies adopt AI-enabled automation, content generation, and efficiency and report encouraging data showing accelerated document generation, predictive insight, and real-time compliance management, these advances in adoption also expose a widening readiness gap. Model governance, good practice validation, and data integrity standards have not evolved as quickly as the technology itself, leaving many organizations unsure about how to adopt AI safely, effectively, and at scale.3To better understand the state of the industry, the Regulatory Affairs Professionals Society (RAPS) and PwC collaborated to benchmark how regulatory functions are engaging with and investing in AI in 2025. The survey aimed to identify current and future areas of AI investment, highlight implementation challenges and successes, and explore factors driving adoption hesitancy. The survey also focused on six key investment areas to build a holistic understanding of investment in regulatory AI and capabilities within the regulatory divisions of strategy, operations, intelligence, compliance, advertising promotion, and labeling.
The survey captured insights from more than 660 participants across 60 countries, spanning pharmaceuticals, medtech [medical technology], healthcare, contract research, and regulatory agencies. Respondents represented a spectrum of company sizes from early-stage innovators to multinational leaders, with roles ranging from technical contributor to senior executive. Together, their perspectives offer a unique, global view of how regulatory AI is being explored, tested, and embedded across the regulatory ecosystem. To supplement the data, PwC conducted interviews with thought leaders, adding nuance to quantitative trends and revealing how cultural and organizational dynamics shape adoption.
Regulatory teams recognize the potential of AI but remain measured in their implementation, balancing their ambitions with accountability. Momentum for adoption is building, exemplified by the 45% of respondents investing in regulatory AI and 68% planning to invest in the next two years. As the field moves from experimentation to execution, success will depend on more than technology. Successful adoption will require investment in governance, collaboration, and human capability while validating perceived value against measurable impact. This article explores this evolution – how regulatory affairs can transition from managing compliance to leading innovation, enabling AI to become a trusted partner in advancing both scientific integrity and public health.
Methods
The survey followed a sequential design and mixed-methods approach, beginning with qualitative, exploratory questions that informed the subsequent quantitative phases. The survey questionnaire included a mix of single-response, multiple-answer, and free-response questions along with interviews over videoconference, designed to capture both quantitative and qualitative insights into regulatory AI adoption. For single-response and multiple-answer questions, results were aggregated and reported using absolute values to reflect the total number of selections across respondents. Free-response questions were analyzed independently through qualitative review to identify recurring themes, notable patterns, and illustrative examples.In cases where respondents selected Other for a multiple-choice question, the free-text responses were reviewed and, where appropriate, reclassified into existing predefined categories to enable analytical consistency and reduce ambiguity. This approach maintained data integrity while allowing for flexibility in capturing novel or emerging perspectives not initially included in the programmed survey responses.
To explore how organization size influenced adoption of regulatory AI, companies were categorized by employee count: small (1-99), medium (100-999), and large (≥1,000). This classification enabled comparative analysis across tiers, revealing how company size correlated with readiness and implementation strategies, and how it influenced adoption and readiness.
Results and findings
Use case adoption and complexityAs regulatory environments grow increasingly complex and data intensive, organizations are turning to AI to enhance efficiency, accuracy, and agility. Nearly half of the survey respondents said they had invested in regulatory AI solutions (Figure 1), signaling that regulatory functions are moving beyond experimentation and toward adoption. In contrast, nearly half of respondents reported their organization has not yet invested in AI, showing that there is still hesitation to implement new tools (see Challenges and Hesitation section). This investment exemplifies the strategic importance of AI-driven intelligence and adaptive technologies in managing regulatory submissions, responding to evolving global requirements, and accelerating the time to market.
The survey also examined adoption trends by company size and type, revealing meaningful differences in how organizations approach adoption of regulatory AI. Larger companies are leading the way in adoption because of their greater access to capital and established digital transformation roadmaps. Many have already embedded AI within enterprise-level regulatory systems with measurable efficiencies and compliance gains. In contrast, small and medium-sized organizations are taking a more cautious, “wait-and-see” approach, closely monitoring the RoI and proven outcomes from early adopters before committing significant resources.
Responses were also examined by the following categories of company type: research and education, healthcare services, media and communications, pharma/life sciences and medtech, and independent/other (i.e., nonprofit, self-employed). Adoption was broadly consistent across segments, indicating that the drivers of adoption are industry wide rather than sector specific. This even distribution suggests that the perceived value of AI in regulatory functions is universally recognized, even if the pace and scale of implementation vary by organizational maturity and resources.
Among organizations that have adopted regulatory AI, applications are diverse but aligned with off-the-shelf, high-value use cases (Figure 1). The most common area of implementation reported was strategic intelligence (65%), where organizations are leveraging AI-driven strategic intelligence platforms to continuously monitor global health authority updates, analyze competitor activity, and anticipate policy shifts that impact submission strategies. Chatbots for research represented the second most common use case (59%). AI-powered chatbots are being deployed to support regulatory researchers by rapidly retrieving guidance documents, summarizing regulatory precedents, and answering complex compliance-related queries in real time. Other highly adopted use cases included drafting of the electronic common electronic document (eCTD) and quality control (50%) and health authority question-and-answer support (49%). This reflects a focus on using AI to unlock efficiencies in information management, reduce manual workload, and improve the consistency and compliance of regulatory outputs.
Areas where adoption is lower, such as labeling and leaflet generation (32%), promotional claims substantiation (23%), compliance management (22%), and other niche applications, tend to have greater operational and technological complexity and carry higher regulatory risk. These processes involve multiple layers of cross-functional review, system integration, and sign-off with strict compliance dependencies and external consequences if errors occur. For instance, labeling and promotional materials require precise alignment with global regulatory standards and marketing claims, while compliance management tools should integrate with validated systems under strict audit controls. In addition, many of these activities occur within secure environments that require authenticated log-ins and traceable workflows, making AI deployment more technically challenging. As a result, organizations are proceeding cautiously, prioritizing AI adoption in lower-risk areas before scaling to functions with greater compliance exposure.
Adoption maturity
The survey prompted respondents to assess their organization’s overall level of AI maturity within regulatory functions. These questions were aimed at ascertaining whether AI tools are being used and how deeply they are embedded within organizations. The results show a broad distribution across maturity levels showing a concentration of organizations in the early to intermediate stages of adoption (Figure 2). Many respondents indicated that their AI initiatives remain in pilot or proof-of-concept phases, reflecting a continued focus on testing and validating business value. A smaller proportion reported more advanced integration, which is characterized by cross-functional implementation, automated workflows, and established governance structures. The variability in maturity levels indicates that while industry-wide progress is evident, full maturity in regulatory AI remains an emerging frontier.
The responses indicate that most organizations remain in the early stages of regulatory AI adoption, with 44% reporting activity in the proof-of-concept or pilot phase, 39% in active implementation or scale-up, and only 17% achieving full implementation. Among the top use cases, strategic intelligence and eCTD document drafting and quality control stand out as areas of concentration due to the availability of off-the-shelf solutions that lower technical barriers and enable quicker deployment. Large organizations dominate the pilot and implementation phases, leveraging structured innovation pipelines and resources to test emerging technologies. Of note is that small and medium-sized companies, while fewer in number overall, account for a greater share of full implementations. This stems from their organizational agility, streamlined governance, and less complex legacy systems, which allow them to move faster from pilot to operational use while also often implementing more sophisticated, end-to-end regulatory AI solutions with fewer change management hurdles.
Overall, the findings depict a regulatory landscape amid digital transformation where AI adoption is advancing but full maturity has yet to be realized. Organizations are prioritizing accessible, off-the-shelf products for use cases such as intelligence gathering and eCTD drafting to gain early efficiencies while mitigating implementation risk. Larger companies continue to lead in piloting and scaling solutions, supported by greater resources and established digital infrastructure. In contrast, smaller and medium-sized organizations are waiting to invest, although some are ?already? demonstrating notable success in achieving full implementation through agility and streamlined change management. Together, these patterns suggest that although regulatory AI adoption is still evolving, momentum is building across company sizes and sectors, signaling a steady shift toward more integrated, intelligent, and data-driven regulatory operations.
Intended business value and leading practices
As part of the benchmarking survey, respondents were asked to identify the primary business objectives driving their regulatory AI initiatives. These questions sought to uncover how organizations prioritize value and the key objectives and business problems attempting to be solved. Respondents reported that operational efficiency is the dominant driver of investment in regulatory AI across all use cases, reflecting the industry’s desire to leverage AI to reduce manual workload, streamline documentation, and accelerate response times (Figure 3). At the same time, adherence to regulatory requirements ranks as a strong secondary goal, signaling that organizations view regulatory AI as a compliance tool as well.
Beyond efficiency and compliance, companies are beginning to target higher-order outcomes, such as cross-functional collaboration, asset strategy, and patient safety, although these remain lower priorities. For example, respondents who are using AI for labeling and leaflet generation report doing so to improve patient safety and data integrity, indicating potential for AI to support end-to-end regulatory and safety alignment. Overall, the data suggests a balanced but evolving mindset: organizations are primarily focusing on adaptive process optimization and beginning to incorporate AI into quality-focused applications. The next frontier will be connecting these efficiency gains with measurable improvements in data accuracy, collaboration, and decision-making quality across the regulatory ecosystem.
As organizations begin to operationalize AI within regulatory frameworks, understanding the primary enablers of successful adoption becomes increasingly important for prioritizing investments and organizational change efforts. The survey results reveal that access to high-quality datasets stands out as the most frequently cited driver of regulatory AI success across all use cases, identified by 29% of respondents (Figure 4). This finding highlights that the effectiveness of AI in regulatory settings is heavily dependent on the availability of reliable, representative, and well-curated data, which becomes increasingly important as models are fine-tuned for context-specific decision making. Without this foundation, even advanced AI models risk producing outputs that lack regulatory validity or interpretability.
Close behind access to high-quality datasets, leadership support and sponsorship (26%) emerged as a crucial determinant of progress. Executive endorsement not only secures funding and resources but also fosters a culture of innovation and risk tolerance that is essential for piloting AI in traditionally conservative regulatory environments. Similarly, the availability of technical datasets (26%) reflects the growing recognition that AI systems require both operational and experimental data sources to deliver actionable insights across diverse regulatory workflows. Finally, cross-functional collaboration (18%) remains a vital but evolving success factor. Many organizations appear to be in the early stages of bridging the gap between technical AI teams and regulatory subject matter specialists, a necessary step for building trust, interpretability, and long-term sustainability. Collectively, these insights reinforce that although data quality forms the cornerstone of regulatory AI success, leadership engagement, interdisciplinary collaboration, and governance maturity are equally essential for translating technical potential into regulatory value.
Challenges and hesitation
Despite growing enthusiasm around regulatory AI, many organizations face persistent challenges that hinder widespread adoption and cause hesitation in investigating potential technology options. The most significant barriers cited by respondents include trust in technology, the ability to measure success, and budget constraints (Figure 5). Many regulatory teams remain cautious about relying on AI outputs for critical compliance decisions, citing concerns over accuracy, data provenance, and interpretability. Without a clear understanding of how AI systems arrive at conclusions, organizations often struggle to build confidence in their use, especially in the highly regulated industries where accountability and precision are paramount.
Financial and operational limitations compound these challenges, creating a significant barrier to scaling AI within regulatory environments. Many organizations, particularly smaller or medium-sized firms, often struggle to allocate sufficient funding for AI initiatives amid competing business priorities such as compliance modernization, digital transformation, and resource optimization. Budget constraints often lead to fragmented or pilot-level investments, limiting the ability to build the sustained infrastructure and governance models required for regulatory AI. At the same time, a persistent shortage of specialized interdisciplinary talent, including data scientists with regulatory knowledge and regulatory professionals with AI literacy, further slows implementation, creating dependency on external vendors that can inflate costs and dilute institutional learning. Even when leadership recognizes the potential long-term efficiency and compliance benefits, the absence of standardized frameworks for quantifying RoI in regulatory contexts presents a major obstacle. Traditional financial models struggle to capture the intangible gains from improved data traceability, accelerated submissions, or enhanced audit readiness. As a result, many organizations default to a wait-and-see approach, deferring investment until clearer value metrics, regulatory guidance, or peer success stories emerge. This cautious stance, while fiscally conservative, risks widening the gap between early adopters and lagging organizations, which ultimately reinforces an uneven maturity curve in the adoption of regulatory AI across the life sciences sector.
Data privacy and technical maturity also pose significant risks, with many reporting persistent fears of exposing proprietary data or inadvertently breaching compliance obligations remain strong deterrents, particularly when organizations rely on external or cloud-based AI vendors. Regulatory teams often manage highly sensitive information regulated under intellectual property or patient privacy laws, making even perceived vulnerabilities unacceptable. These concerns are compounded by inconsistent global data protection standards and the growing complexity of cross-border data transfers, which introduce uncertainty about jurisdictional compliance and liability. On the technical side, respondents are concerned that many AI vendors lack deep regulatory domain expertise, resulting in solutions that fail to reflect the precision, traceability, and auditability required in a regulated environment. The combination of these factors not only heightens operational risk but also contributes to a broader trust gap between regulatory professionals and AI technologies. These concerns reinforce the need for more transparent, domain-aligned, and secure AI ecosystems.
Beyond technical and structural challenges, human concerns play a critical role. Regulatory professionals, accustomed to reducing risk and establishing compliance, are often hesitant to embrace AI’s iterative and experimental nature. Anxiety about job displacement, particularly among medium-level documentation and quality roles, further amplifies resistance. Building trust in technology, investing in upskilling, and fostering transparent governance will be essential for advancing the adoption of regulatory AI. The industry’s ability to address these intertwined technical, financial, and human factors will help determine how effectively AI reshapes regulatory functions.
Industry insights and readiness
The current regulatory landscape reveals an industry still in the early stages of realizing the full promise of regulatory AI. While organizations are actively piloting use cases, most have yet to capture its enterprise-wide potential or demonstrate scalable impact. The survey findings indicate that near-term RoI remains a dominant priority, with many companies under pressure to show rapid, measurable results that justify investment in AI and maintain executive support. The short-term focus emphasizes the need for demonstrable business value in an economic environment where capital allocation is being scrutinized across all functions. At the same time, AI adoption brings both optimism and apprehension, with the former demonstrated through early successes exemplifying AIs potential to improve accuracy, efficiency, and compliance and the latter, through persistent concerns around workforce disruption, implementation complexity, and long-term sustainability. The balance between opportunity and caution defines the current regulatory environment, raising an essential question: In a rapidly expanding AI landscape, how can you and your organization prepare for the changes and challenges ahead?At the organizational level, preparing for the accelerating integration of AI into regulatory and operational processes requires strategic investment in both capability and culture. Companies should begin by hiring or upskilling talent with AI fluency who can bridge the gap between technical innovation and regulatory rigor while also partnering with industry specialists to guide the transition into an AI-embedded enterprise. Equally critical is the establishment of robust data governance frameworks, confirming that the data feeding AI models is high-quality, traceable, and compliant with privacy standards. Organizations can also invest in structured AI education programs, which equip regulatory, quality, and compliance teams with foundational knowledge to understand, question, and guide AI use responsibly. In addition, early partnerships with trusted advisors can accelerate learning and reduce risk, allowing firms to pilot AI tools in controlled environments before broader adoption. Above all, leadership should set the tone for a responsible, transparent, and adaptive AI culture, which balances innovation with accountability and public trust.
For individuals, preparation begins with curiosity and proactive engagement. The benchmarking results show a growing recognition that regulatory professionals are not being replaced by technology but redefined through it. As AI absorbs repetitive tasks, professionals are shifting toward analytical, interpretive, and strategic roles that demand both regulatory expertise and digital fluency. Professionals at all levels can start by exploring accessible tools such as AI-driven productivity platforms or language models to understand their practical potential within their own roles. Learning how AI might increase individual capacity to focus on strategic priorities, improve documentation quality, or support decision making allows individuals to position themselves as early adopters and change enablers rather than passive observers. When implementing or testing AI systems, it is essential to elevate risks and anomalies quickly, enabling transparency and safety remain top priorities. In addition, fostering cross-functional collaboration between IT, data science, and regulatory experts enables more resilient and informed decision making. Ultimately, those who combine domain expertise with digital literacy will be better equipped to navigate and shape the evolving landscape of AI in the regulatory field, enabling both compliance and innovation to thrive together.
Ultimately, the convergence of regulatory AI, evolving compliance expectations, and industry performance data suggests that regulatory industries are now at a pivotal moment. The organizations that align technological innovation with ethical governance and human capability can set the standard for regulatory excellence in the coming decade. In addition, successful organizations should approach regulatory AI with intention by balancing innovation with accountability and treating it as an evolution of regulatory science rather than a shortcut. The industry is transitioning from reactive compliance to intelligent, predictive oversight, an evolution that promises not only operational efficiency but also enhanced public trust and patient outcomes.
Conclusion
The integration of regulatory AI within regulatory industries signifies more than a technological shift. It represents a fundamental reimagining of how healthcare organizations interpret, manage, and act on regulatory information. The industry benchmarking reveals a sector steadily moving from experimentation to structured implementation. Large companies, equipped with formal AI strategies and stronger digital infrastructures, are realizing measurable gains in efficiency and accuracy. Smaller and medium-sized organizations, while progressing more cautiously, are leveraging partnerships and modular tools to remain competitive in a rapidly evolving regulatory environment. Together, these trends illustrate a maturing ecosystem in which AI is becoming not merely an enhancement but an expectation.Early adopters of regulatory AI are realizing measurable benefits in efficiency, quality, and insight generation, but broad adoption is yet to be observed. The organizations advancing most effectively are those approaching regulatory AI with intention, by balancing innovation with accountability and viewing AI not as a shortcut, but as a disciplined evolution of regulatory science. To continue momentum, regulatory leaders should establish clear RoI measures, invest in governance structures that facilitate data integrity, and foster collaboration between regulatory, data, and quality functions. Partnerships across industry, academia, and technology can help accelerate validation and adoption by confirming new tools maintain compliance and build public trust.
At the individual level, the conversation around regulatory AI is as much about capability as it is about courage. Many professionals worry about AI displacing their roles, yet the opposite is often true with those who understand and apply these tools becoming more valuable. The regulatory workforce of the future will be defined by those who can harness AI responsibly by interpreting insights, refining models, and establishing alignment with ethical and scientific standards. Embracing continuous learning and engaging directly with emerging technologies can guide professionals to transform apprehension into opportunity. The future of regulatory affairs will not be led by machines, but by people who know how to work with them, ultimately turning regulatory AI from a source of uncertainty into a driver of trust, innovation, and lasting impact.
About the authors
Emily M. Wilts, PhD, RAC, is a senior associate in PwC’s business operations practice, with more than nine years of medical device and pharmaceutical experience. During her time at PwC, she has advised clients on R&D operations strategy, technology transformation, and integrating AI into workflows. Wilts earned a PhD from Virginia Tech, working closely with industry to study and develop personalized medical devices, and completed a postdoctoral fellowship at the University of British Columbia, developing insulin-producing tissue patches for diabetes treatments. She holds the Regulatory Affairs Certification in medical devices and is a RAPS member. She can be reached at [email protected]
Numi Prasad, BS, is a manager in PwC’s Health Policy & Intelligence Institute practice, with more than eight years of experience in healthcare and pharmaceuticals. During her time at PwC, Prasad has advised clients on impacts from legislative developments and regulatory changes, global regulatory strategies and submission plans for pharmaceuticals, medical devices, digital health, and combination products. Before joining PwC, Prasad had various regulatory roles ranging from regulatory strategy and QMS management to developing safety processes for audit readiness. She earned a bachelor of science degree in cell biology and molecular genetics from the University of Maryland and is a RAPS member. She can be reached at [email protected]
Denise Fulton, BS, is vice president for research & content strategy at the Regulatory Affairs Professionals Society, where she leads the Regulatory Focus, journals, books, and research teams. She joined RAPS in early 2020, having spent 20 years leading newsrooms for MDedge/Medscape, a company dedicated to covering medical news for physicians and other healthcare professionals. She has an additional 10 years of association experience. She can be reached at [email protected]
Mike DeMarco, PharmD, is a director in PwC’s risk and regulatory practice with more than 15 years of healthcare and pharmaceutical experience. During his time at PwC, DeMarco has advised regulatory clients on differentiating capabilities strategy, global operating model designs and technology transformation. Before joining PwC, he was a regulatory strategist, leading health authority engagements and developing Phase I-III global regulatory strategies focused on cardiovascular-metabolic drug development for novel compounds. DeMarco earned a doctorate in pharmacy at Butler University and completed a postdoctoral program in regulatory strategy, rotating between industry, academia, and the FDA’s Center for Drug Evaluation and Research. He can be reached at [email protected]
Disclaimer ©2025 PwC US. All rights reserved. PwC US refers to the US group of member firms and may sometimes refer to the PwC network. Each member firm is a separate legal entity. See www.pwc.com/structure for further details. This content is for general purposes only and should not be used as a substitute for consultation with professional advisers.
Acknowledgments This article was adapted from a presentation by the authors at the 2025 RAPS Convergence in Pittsburgh from 7-9 October.
The authors thank Dhruv Mansharamani, Ella Berg, Erik Gunther, Jeffrey Jeyakumar, Niranjana Unni, and Zhi Chai for survey development, data analysis, and editorial support.
Citation Wilts E, et al. Potential to performance: How regulatory organizations are adopting AI. RF Quarterly. 2025;5(4):45-55. Published online 15 December 2025. https://www.raps.org/news-and-articles/News-Articles/2025/12/Potential-to-performance-How-regulatory-organizati
References
References were last checked and verified on 7 December 2025.
- Guleria S, et al. Artificial intelligence integration in healthcare: Perspectives and trends in a survey of US health system leaders. BMC Digit Health. Published 19 November 2024. Accessed 11 November2025. https://doi.org/10.1186/s44247-024-00135-3
- Poon GE, et al. Adoption of artificial intelligence in healthcare: survey of health system priorities, successes, and challenges. J Am Med Informatics Assoc. Published 5 May 2025. Accessed 11 November 2025. https://doi.org/10.1093/jamia/ocaf065
- Lamberti MJ, et al. The adoption and use of artificial intelligence and machine learning in clinical development. Ther Innov Regul Sci. Published 29 May 2025. Accessed 11 November 2025. https://doi.org/10.1007/s43441-025-00803-0
- DeMacro M, Fulton, D. From potential to performance: How regulatory organizations are adopting AI. Presented at RAPS Convergence; 8 October 2025.
