Unsustainable regulatory workloads leave no choice about AI adoption
To determine where pharma/biopharma regulatory functions see the greatest potential for AI technology, ArisGlobal commissioned an exclusive survey1 of senior regulatory professionals in US pharma and biopharma organisations. The research was conducted in early September 2024. Discussing the findings, Renato Rjavec from ArisGlobal notes that, although plans to invest in regulatory AI solutions are high, many companies are standing in their own way of progress.
Life science companies have no choice but to be smarter about how they allocate resources. That’s if they want to maximise affordable access for patients and maintain commercial viability. It is why the industry is turning purposefully now toward artificial intelligence (AI), and in particular next-generation technologies such as Generative AI (GenAI) powered by large language models (LLMs).
In many cases, however, life sciences companies are standing in the way of their own progress.
Process inefficiency is out of control
That regulatory workloads are soaring across the international life sciences industry is well understood. In our new survey, with senior regulatory professionals in US pharma and biopharma organisations, almost all (97% of) respondents had seen their regulatory obligations expand over the last five years, with three in five (60%) putting the increase beyond what would be expected as the result of company growth. Expectations are that this growth will continue over the next five years.
The individual regulatory process challenges are multiple, including excessive time spent producing submissions/dossiers; maintaining labelling compliance; inputting data/documents into IT systems; verifying submission correctness/completeness; performing regulatory impact assessments; and locating data or documents in existing IT systems. More than a quarter of the research base indicated each of these issues. Further challenges include responding to agency queries; inadequacy of current IT systems; and time lost to administrative tasks (such as data quality checks, and assessing submission readiness).
A lack of qualified people was least likely to be registered as a serious challenge or concern, suggesting that preferred strategies do not involve allocating more people to processing regulatory workloads. Rather, regulatory functions are looking to smarter use of technology.
AI’s perceived value to the regulatory function
There is now widespread acceptance of AI’s potential usefulness in solving information or process bottlenecks in a regulatory context, with 96% of survey respondents citing its current or potential value here, and almost half (45%) describing AI as “very useful”. Only 4% said they couldn’t see any value for the technology for this purpose.
By use case, almost all respondents could see direct potential for AI in transforming labelling compliance and deviations maintenance; capturing, searching, filtering the latest regulatory requirements; automating the intake of Health Authority interactions; automating regulated content translations for different markets; automating the authoring of responses to Health Authority queries; suggesting improvements to submissions/dossiers; performing regulatory impact assessments; authoring submission documents; automating document summarization; and generating entire regulatory submissions.
Although specialist AI tools and applications for targeted regulatory use cases are only now coming to market, over a third (35%) of respondents claimed to be using AI for regulatory purposes in some form already, while 42% plan to invest in the next 18 months. A further 15% are looking at a timeframe beyond that, but do also have plans to roll out AI within the regulatory function.
Outdated mindsets are standing in the way of progress
Although the appetite to embrace AI appears markedly high, there are clear barriers holding companies back.
Asked what would prompt their company to invest/expand investment in regulatory-focused AI capabilities now or in the near future, respondents most commonly cited the discovery that their competitors are using the technology (41%). Other triggers would be soaring workloads/continued resource pressures (40%); the technology becoming more mature and proven (36%); the availability of specific tools geared to their problem tasks (35%); and relevant IT systems becoming easier and more affordable to deploy (33%).
Endorsement or recommendation of AI by regulators would inspire investment also for just under a third of respondents.
Budget did not appear to be a barrier to investment plans: just 18% indicated that the availability of new budget would unlock AI investment. That budget constraints are not a major barrier is encouraging, because hesitancy linked to “a lack of confidence to deploy” is surmountable.
Certainly AI technology, including Generative AI (GenAI) is maturing and advancing at an accelerating pace, and specialist applications for target use cases in a life sciences regulatory context are being actively developed and piloted today, showcasing what is possible. This is in keeping with Gartner’s prediction that, by 2027, more than 50% of the GenAI models used by enterprises will be specific to either an industry or business function, up from just 1% in 20231.
Matters of AI maturity & confidence
Asked what, if anything, was holding back initial or further investment in AI for Regulatory purposes, respondents most commonly cited outdated existing IT landscapes (45%); a belief that risks currently outweigh the benefits (44%); and inadequate availability/quality/consistency of data or content resources to derive the value from AI (42%).
In addition, 39% of respondents felt the technology remained too immature/unproven; similarly, that the tools do not exist today to address their particular regulatory pain points. Sixteen per cent blamed a lack of trust in AI currently. This was ahead of budget challenges: only 15% named a lack of budget as a barrier to AI investment.
Regulatory AI aides feature strongly in the future
Finally, the research sought a sense of respondents’ expectations of AI in a regulatory context over the longer term, asking which of several scenarios they agreed with as being most likely in the future. Almost half (48%) of respondents agreed that, in time, AI would transform a lot of routine regulatory work and considerably streamline processes. Over 2 in 5 (43%) felt AI would drive up accuracy and quality in the information they produce for regulators and patients. Almost 2 in 5 (39%) respondents believed AI would be critical to the regulatory function’s ability to keep pace with market demands. And over a third (35%) of respondents agreed that AI would save a lot of time and money.
This supports the finding above that pain points are multiple and diverse, and suggests that ideally an investment in the right AI capability should ultimately enable all of these to be targeted.
Recommendations
Awareness of next-generation technologies like GenAI, and their potential to transform high-intensity routine workloads, is high now. Regulatory functions, along with other departments and executive boards, well appreciate the potential to streamline processes and make them tighter and more consistent.
As the technology continues to mature, and as specialist GenAI solutions become available that target specific regulatory pain points with demonstrable benefits, the real yet surmountable barrier to progress remains one of “how”. How can companies take advantage of AI? How can they deploy it within their existing IT estates? How will they know how to use it, and that they can trust it?
To cultivate that familiarity and confidence, finding a targeted use case to test what the technology can do is a practical way forward. Taking an agile, incremental, use-case-by-use-case approach to GenAI deployment will be faster, and represent lower cost and lower risk, than a big-bang “AI project”. It is also more likely to build engagement, as specific incremental wins are demonstrated.
Whether companies are exploring AI in earnest for the first time, or looking to increase the technology’s usage and impact, however, they must first get their data in order and stabilise their core systems, so that they can then fully streamline and optimise their processes with AI’s help.
About the research
This report is based on an exclusive survey conducted for ArisGlobal by Censuswide between August 30 and September 06, 2024. The research was conducted among a sample of 100 US respondents with senior titles in regulatory roles within Pharma and BioPharma companies (25+).
References:
1: https://www.arisglobal.com/resources/regulatory-industry-survey/
About the author
Renato Rjavec is Senior Director of Product Management at ArisGlobal, where he is shaping the future of regulatory information management as well as quality management for life sciences, with a keen focus on AI as a means for targeted automation of critical but labor-intensive processes where accuracy and precision are paramount. Renato has almost two decades of experience in ideation, development and implementation of regulatory and quality solutions for the life sciences industry.
