Q&A: PBMs’ Adoption of AI Raises Efficiency Hopes, Ethical Concerns

As pharmacy benefit managers (PBMs) begin leveraging AI to streamline everything from claims processing and prior authorizations to formulary management and fraud detection, the pharmacy community is left asking a question that cuts to the heart of an already complicated relationship: Will AI make PBMs better partners?

In this exclusive interview, pharmacist and pharmacy technology expert Timothy Aungst, PharmD, offers a measured look at what AI integration actually means for the PBM space, separating genuine operational promise from the ethical and transparency concerns that are likely to follow.

“These questions, I suspect, will increase as clients become more aware of AI intricacies and what it may mean for their members,” Aungst told Drug Topics®. “I suspect we’ll see some developments down the road where we establish core ethical principles and governance on these matters, but for now, I think there’s a lot of questions, especially on risk and data protection, that are going to be asked during the interim.”

For pharmacists already skeptical of PBM practices, this conversation provides important context for understanding how AI could either improve or further complicate an industry relationship that has long been defined by tension.

READ MORE: How Will Emerging AI Impact the Pharmacy Benefit Manager Space?

Drug Topics: From your expertise and experience in the pharmacy industry, what are your initial thoughts when you hear AI technology spilling into the operations of pharmacy benefit managers (PBMs)?

Timothy Aungst: Like every other industry, especially those whose processes have a lot of data to process and utilize, I think it was inevitable that PBM and related businesses in the pharmacy landscape would look to AI as a means to help with their operations.

Now, the caveat is that I would also argue this isn’t new, and we’ve seen previous work from PBMs and adjacent organizations leveraging technologies like business intelligence and machine learning for certain items. This new AI wave is likely driving a stronger push toward newer capabilities that may not have been achievable with earlier technologies, and thus a larger investment of time and money is going on I suspect.

Drug Topics: In what ways do you see AI coming in and truly improving the pharmacy benefits industry?

Timothy Aungst: This, I think, will have to be seen as to how we define what improvements to quantify for the industry. If it’s simple daily operations, I think most may see that benefit from the integration of certain AI technologies (eg, large-language models, GenAI) into workflows for certain daily tasks that their workforce may embrace. The crux will be the issue of how using AI tools may yield a change in their core business operations, whether that is streamlining prior authorizations (PAs) or claims, improving interactions with members, identifying fraud, aiding in contracting, formulary management, clinical services, or other key business arms.

I think this will be the interesting thing to watch in the coming years. I suspect some are low-hanging fruit due to a data-driven and process-oriented culture that AI could be adapted to due to established workflows and may help seed up those processes (eg, claims processing, PAs), but others may need more experimentation to see if the current AI technology can be of service for the PBM and their clients.

Drug Topics: Ethics seem to be a key concern regarding the AI-PBM phenomenon, within a pharmacy benefits industry typically known to involve predatory and “unethical” practices—for a lack of a better word. What are your ethical concerns regarding PBMs’ use of AI for their practices and operations, and how may these concerns affect patients as AI is more broadly adopted?

Timothy Aungst: I suspect the biggest questions clients will raise, especially those using PBMs that leverage AI for ops, are about how that may impact them. How is personal health information (PHI) and data handled? Who owns the risk to their members if the AI model becomes biased or leads to unintended results that a human may have caught? How much power is the PBM entrusting to its models to take on with decision-making?

These questions, I suspect, will increase as clients become more aware of AI intricacies and what it may mean for their members. I suspect we’ll see some developments down the road where we establish core ethical principles and governance on these matters, but for now, I think there’s a lot of questions, especially on risk and data protection, that are going to be asked during the interim.

Drug Topics: With PBMs being a major buzzword of negativity in pharmacy—as well as AI seemingly becoming a buzzword in general society today—what do you think pharmacists will think about AI use among PBMs?

Timothy Aungst: I suspect, similar to other industries, it’ll be mixed. Am I dealing with a human? Who made these decisions? When can I talk to someone who will help me with a patient need that wasn’t anticipated? How do I navigate all of this?

These are questions I think pharmacists will start raising. If any company wants AI to become a part of their ops and have positive engagement, a level of trust and transparency will be needed. I can see that some may do this well and others may have difficulty with rollout and may find that AI works in some areas and gets buy-in from the pharmacy community, and others will get massive pushback.

PBMs will have to figure that out if they hope for AI to deliver on its theoretical benefits to a practical market.

READ MORE: Technology and Data Resource Center