Can AI improve pharma compliance?

Compliance expert agrees that artificial intelligence (AI) can enhance compliance functions, but it will require careful implementation and human oversight to realize its full potential.

Because the world, including how pharmaceutical companies operate, communicate, and serve patients, is rapidly evolving, compliance efforts must adapt, enhance, and leverage the tools available, according to Audrey DeGuarde, vice president, customer success and compliance operations, at RLDatix, who spoke on a panel during the “Pharmaceutical Compliance: Emerging Trends & Best Practices for 2025” webinar, hosted by the Pharmaceutical Compliance Congress in March.

Audrey DeGuarde, vp, customer success and compliance operations, RLDatix

Data, for example, is central to compliance, she noted. “It’s certainly a core part from a compliance perspective of making sure we're collecting that data and then using that to support us,” she said.

“When I think about AI, how does AI layer in to help solve that problem? We need to crawl before we walk or run,” she continued.

Results from a live poll during the webinar showed that about half of attendees said their company was exploring and testing AI in small cases.

“We’re talking about it in a very theoretical way. What can we do? How can we do it? Maybe testing it in some kind of pilot scenario, in some part of our business. I think it’s because the AI conversation raises a lot of risk questions about how it interplays with the human element,” she said.

DeGuarde highlighted three immediate, practical AI applications for companies’ compliance efforts, noting that scalability is a big one.

“We are talking about volumes [of data]. We are all thinking about how we can collect data and use it in various capacities from a compliance perspective — a lot of data for what we have to go through and look at to try and tease out things that might highlight risk for us,” she said.

She explained that AI can process large amounts of data, finishing tasks in hours or days compared to the weeks it would take manually. She further stated that AI should handle the laborious process of sifting through data to extract key information.

The second benefit is risk spotting. After AI processes large datasets, it can identify risks by recognizing patterns in data, emails, correspondence, and behaviors, flagging variances that require attention. “Pattern recognition is an easy way to implement AI,” she said.

The final approach DeGuarde discussed was real-time reaction. “If we are combing through volumes of data more quickly and using it to identify patterns that might be concerning to us, it puts us from a compliance perspective, in a position where we can react in real time, or near real time, to items that might be a concern or might be a risk,” she said.

However, human oversight will still be needed, she added. “We still need to put our eyes on it to say, is this, in fact, a risk that we're concerned about.”

AI is “enabling us to get in front of those concerns more quickly, allows us to pivot and provide feedback so the business can continue doing what needs to be done, and allows us to make sure that if there are concerns … stopping them and preventing them from being replicated over time,” DeGuarde said.

Unsplash/Numan Ali