Generative AI – Unlocking Opportunities In Regulatory Affairs

A Vision For Streamlined Processes And Enhanced Efficiency

Within two months of its launch, ChatGPT had already registered 100 million users setting a record for the fastest growing base of users ever. Such unprecedented adoption demonstrates the potential that decision makers in many industries see in generative AI. The use of this technology in the life sciences sector can drive efficiencies, unlock innovation, improve compliance oversight, and ultimately facilitate more rapid access to medicines for patients who need them. ?

We believe that there are several powerful use cases in the Regulatory Affairs domain that will demonstrate the capabilities of generative AI in driving process efficiencies, reducing redundancies, using enterprise data more effectively and ultimately, streamlining the end-to-end value chain for Regulatory Affairs organizations. This will, in turn, unleash resources to focus on more complex activities much needed as complexity increases due to the growth in exploration of new medical frontiers including novel therapies and the overall increase in focus on innovation.

The Pharma regulatory affairs domain is one of the most regulated areas across industries. The regulatory affairs organizations are a key part of the value chain and have a multitude of co-dependencies with other organizations.? Whilst the upstream input from a regulatory affairs organization is strategic input (to preclinical, clinical, and CMC strategies and more) the most obvious downstream output is the regulatory dossier (including all its sections from modules 1 to 5).?

We believe that there is a compelling case to use Generative AI to (semi) automate the authoring of regulatory documents across of the modules 1 to 5. Most regulated countries follow the eCTD (Common Technical Document) structure. Whist the high-level structure/ skeleton is standardized, the expectations for the narrative/ text within the respective modules will vary accordingly to various aspects including therapeutic area and geography to name a few. All the above needs to be considered when developing a value adding solution, making the challenge of automating the authoring of regulatory modules a highly engaging and worthwhile challenge. Whilst there has been much hype around the automation of authoring of modules for many years, the investment dollars did not follow, and little true progress has been made. Given the power of Generative AI, we believe that the playing field has changed, and the time is right to reinvigorate efforts related to automation of authoring.?

Another use case we think should be explored is that of the gathering and interpretation of regulatory intelligence. Whilst the top line structure of regulatory dossier is somewhat standardized, as mentioned above, the local and regional regulatory requirements are far from standardized. One of the main concerns of COOs of Pharma companies today is the inability of the organizations to keep up with the changing regulations, and the rate of change is accelerating. Lack of understanding of the regulations can lead to delays in market entry and in the worst case to compliance breaches with direct financial impacts. The use of Generative AI to extract regulatory intelligence from directives and regulations directly and subsequently summarize and format in a user-friendly fashion can lead to efficiencies and a dramatic increase in reliability of the information. This will certainly reduce time to market which will have a direct positive impact on patients in need and drive efficiencies for the organization.

Another area where Generative AI will almost certainly add tangible value is the area of optimizing the use of enterprise data. Some say that data is the oil of the 21st century. Data stewardship and data integrity have become front and center for both industry and the regulators, however, most organizations still have multiple “silos” of data along the value chain from research to development, to in market products. For example, few companies today have managed to streamline the data flow between regulatory affairs, quality assurance and supply chain and don’t have a joined-up R&D data strategy. Generative AI can be used to support the creation and interrogation of easy to interpret Master Data Sets/ data lakes leading to more efficient, effective, and reliable use of data across the value chain. This will lead to increased productivity and reduce the chance of errors (hence reducing compliance risks).

The above are three illustrative use case and there will be many more as the technology evolves.

Given that Pharma is one of the most regulated industries, we believe that companies will have to “walk before they run”. Early wins in the form of credible and successful pilots will be a key success factor as this will instill confidence across the organization. Early resistance by staff who may perceive Generative AI a threat to their future, should not be underestimated despite the reality being that the effective use of Generative AI will likely reduce the burden on them to do repetitive tasks and allow them to reallocate to more complex, higher value add tasks.? Communication and change management will be a key success factor.?

Trust will be another key success factor as we use Generative AI in the regulatory affairs domain since the sponsors need to trust the accuracy of the data and the relevance of the narratives produced from the data. Likewise, the regulators will need to build trust in the accuracy, integrity and completeness of data being submitted in in the regulatory files, for review.? Given that this is a rapidly developing, and so far, immature area, we believe that it is imperative to have human quality assurance checks in place across the value chain. The regulators will also likely “incentivize” sponsors to maintain such human quality checks until the technology evolves.? Only over time and with maturing of the technology can one consider reducing the thoroughness of such checks.

As Generative AI will be employed to extract data from a multitude of sources and create narratives in the regulatory documents the line between source and regulatory documentation will likely become more blurred than ever. This may present opportunities to further streamline processes, remove redundancies across functions and drive efficiencies across the entire value chain.?

We at Genpact have deep domain expertise in the regulatory affairs space. We also have broad and deep analytics capabilities and are engaged across multiple industries and domains on using Generative AI to drive value for our clients. The combination of deep domain expertise and our strength in analytics give us a unique edge in this space.

He who dares will win.?

This article is authored by Genpact's Global Head of Regulatory CMC Practice, Michael Malone and first appeared in Healthcare Business today.