AI’s Next Step in Pharma Quality: From Detection to Closed-Loop Prevention

According to Shawn Opatka, VP and GM Honeywell Life Sciences, nearly all organizations—about 99%—are engaging with artificial intelligence in some capacity, though the degree and focus of implementation vary widely. Many companies are piloting AI tools within quality, regulatory, and traceability operations to address persistent challenges such as labor shortages, skills gaps, and the overwhelming volume of data inherent to modern pharmaceutical operations.

Opatka emphasized that while AI experimentation is widespread, the industry remains in the early stages of full-scale deployment. Companies are still testing how best to integrate AI into regulated environments, where oversight and validation requirements remain stringent. This experimentation phase, he suggested, reflects a broader willingness within pharma to embrace innovation while maintaining compliance with evolving standards for data integrity and product oversight.

Ultimately, the sector’s cautious but active engagement with AI represents a balancing act: driving efficiency and insight through automation and predictive tools while ensuring transparency and control remain uncompromised. As regulators sharpen expectations around digital accountability, AI’s role in improving data traceability, quality assurance, and operational resilience will only grow more central to the pharma enterprise of the future.

He also dives into the biggest risks companies face if they attempt to scale AI solutions without a strong quality and compliance framework in place; the role of AI evolving within pharmaceutical quality systems; the definition of successful AI adoption in pharma; and much more.

A transcript of his conversation with PC can be found below.

PC: How do you see the role of AI evolving within pharmaceutical quality systems, from reactive issue management toward more proactive assurance and risk prevention?

Opatka: From a reactive and proactive standpoint, we've actually been doing a bit of that for a while now, and have some adoption with our customers. What we would really like to see and focus on, and starting to see some traction is, closing the loop in the process.

You could take a simple, proactive or reactive example where you present a user with some data and allow them to make a choice based on some data that AI has. It's hey, here's what we think happened, and let them make that decision and finally decide, okay, this is the reason why that happened, but using AI to actually assist the operator, but then close the loop on the process to go back and then provide that data in a way that provides more insight and knowledge in a closed loop. That's really where we're focused on, and I think the use cases that are really exciting.