A Conversation with Brandi Stockton and Martin Heitmann on AI in Regulated Pharmaceutical Operations

Bridging the Gap in AI Guidance

As we delve deeper into the realm of artificial intelligence (AI), its integration into the regulated pharmaceutical sector has presented challenges that the industry has yet to fully address. Stakeholders are often left navigating a maze of regulations and guidelines, particularly regarding Good Practice (GxP) environments. This lack of a consistent framework became a significant concern, especially in areas like visual inspection, data analysis, and pharmacovigilance.

To fill this void, the International Society for Pharmaceutical Engineering (ISPE) published the ISPE GAMP Guide: Artificial Intelligence in 2025. This robust resource aims to establish practical guidance for deploying AI in GxP settings, providing a risk-based framework for evaluating, implementing, and maintaining AI systems across the pharmaceutical lifecycle.

Encompassing the Global Landscape

In our discussion with Brandi Stockton, CEO of The Triality Group, and Martin Heitmann, a consultant at the same firm, they emphasized the guide’s comprehensive scope. “Our AI guide encompasses not just pharmaceutical development, manufacturing, and distribution, but virtually every GxP area,” Stockton explained. “It includes AI aspects of medical devices and software-as-medical-device as well.”

The crew behind the guide comprises an international team of experts, ensuring that it reflects regulatory guidance and practical insights from around the world. “Our goal was to support scalable and harmonized implementation across diverse regions,” added Heitmann.

Examining AI Technologies

When discussing the types of AI technologies included in the guide, Stockton and Heitmann illuminated the breadth of approaches covered. “We start with rule-based systems and traditional machine learning, advancing through deep learning and finally to new iterations like generative AI,” Heitmann noted.

This flexibility allows organizations to utilize combinations of these technologies in a way that suits their specific needs. The guide not only promotes a clear understanding of AI applications but also highlights various case studies—ranging from visual inspections of pharmaceutical products to using AI for optimizing biomanufacturing processes and pharmacovigilance tasks.

Navigating Regulatory Expectations

One of the most significant challenges the authors faced was harmonizing regulations across different jurisdictions. “Establishing a clear and widely understood terminology was crucial,” Stockton said. The guide scrutinizes various terminology sources to reach a consensus, especially concerning terms like “validation data set”—which may have different implications in the U.S. FDA context compared to the European Medicines Agency (EMA).

Heitmann highlighted, “While principles like transparency and ongoing monitoring are increasingly standardized, the acceptance of AI applications often varies. Regulatory bodies, like those in the European Union, are beginning to draw clearer lines around acceptable AI uses.”

Lifecycle Management of AI Models

The guide also addresses essential aspects of lifecycle management, emphasizing ongoing quality risk management. “AI models can experience drift and bias over time,” Heitmann pointed out. “The guide promotes a quality by design approach to ensure rigorous monitoring and performance evaluation from conception through ongoing operational stages.”

Both experts underscored the necessity for data understanding. Monitoring input data changes is crucial for maintaining model integrity. “If the data distributions shift, it can significantly affect model performance,” Stockton noted, emphasizing that adjustment protocols are vital.

The Topic of Model Retirement

Chapter 4 of the guide touches on an often-overlooked topic: model retirement. “Model retirement might occur due to inadequate performance or advancements in technology,” Heitmann explained. Factors to consider in this process include traceability of model inputs, outputs, and the relationships with other AI-enabled systems.

Stockton added that planning for retirement should involve an assessment of the residual effects on dependent systems and models. “Using explainable AI methods can assist in understanding the impact of retiring a model, providing a roadmap for informed decision-making.”

Concluding Thoughts

In a landscape rapidly evolving with the integration of artificial intelligence, the ISPE GAMP Guide: Artificial Intelligence emerges as an essential resource. Through thoughtful discussion, Brandi Stockton and Martin Heitmann articulated not only the practical implications of the guide but also the vision behind it—creating a harmonized, risk-based approach that fosters safe, effective AI use in pharmaceutical operations.

This collaborative effort signifies a promising step towards addressing the complexities of modern pharmaceutical practices, propelling the industry into a future where AI operates within a well-defined framework. The guide serves as an invitation for stakeholders to embrace AI responsibly while adhering to the stringent standards that govern their work.

For those seeking to navigate the complexities of AI within GxP environments, the ISPE GAMP Guide: Artificial Intelligence stands ready to illuminate the path forward.

Further Reading:

• ISPE GAMP Guide: Artificial Intelligence, 2025
• ISPE GAMP 5: A Risk-Based Approach to Compliant GxP Computerized Systems (Second Edition), 2022
• Applying GAMP Concepts to Machine Learning, ISPE Pharmaceutical Engineering, 2023
• Machine Learning Risk and Control Framework, ISPE Pharmaceutical Engineering, 2024
• ChatGPT, BARD, and Other Large Language Models Meet Regulated Pharma, ISPE Pharmaceutical Engineering, 2023

About the Experts:

Brandi Stockton is the founder and CEO of The Triality Group, bringing over 25 years of GxP experience to her role. She is an influential voice in the pharmaceutical sector, leading various initiatives to promote AI and quality standards.

Martin Heitmann is a seasoned consultant also affiliated with The Triality Group. With a focus on technology and transformation, he contributes significant expertise in implementing AI within regulated environments.