The Engine of Healthcare: How AI Is Powering Modern PBMs

Demystifying the AI ‘Black Box’ for True Transparency

HR and benefits professionals are experts in crafting competitive benefits packages and ensuring compliance, not engineering. Nowadays, as nearly every vendor and software company advertises their AI add-ons, it can be difficult for non-technical folks to ascertain which companies are actually using AI effectively and not just as a pricey marketing gimmick.

Understanding what to look for in vendors that use AI requires learning the basics of AI terminology. One of the key AI tools that can be used for benefits is a Large Language Model (LLM). These are models that predict the next likely word in a sentence using patterns from massive datasets. This powers applications, like ChatGPT, to generate digestible, contextually relevant text based on their source datasets. 

Healthcare data is notoriously dense, filled with nuances that make it difficult for members to understand. When vendors leverage LLMs, they act as an interpreting layer, taking complexities and transforming them into clear, actionable insights. 

If a pharmacy benefit manager (PBM) has built their systems with modern architecture in mind, they can move beyond the legacy black box of pharmacy benefits. For example, if a member sees "Denied" on a claim without any context, AI-powered Chain of Thought Reasoning could then provide the member with step-by-step explanations of how that status was reached and what to do next. This lends much-needed transparency to the process and reduces frustration for the members and plan administrators.

"Transparency isn't just showing you a number. It's understanding the why. People don’t appreciate black box reasoning. Just the ‘answer’ is probably not going to result in understanding; the step-by-step reasoning probably would."

— Gaston Concilio, VP of Engineering, SmithRx

AI Won’t Replace Humans, It Will Give Them Superpowers

There’s a common misconception that AI is here to replace human experts. At SmithRx, our philosophy is the opposite. We believe it’s a tool that frees humans to do what they do best: empathize and solve problems. Think of it as the difference between using an old hammer and a power drill to build a home; the tool doesn’t replace the builder, it simply allows them to build faster and focus on their craftsmanship. 

By synthesizing data before a support agent even picks up the phone, AI ensures that the conversation starts with a solution rather than a series of “please hold” moments. This Human in the Loop framework ensures that while AI handles the mundane tasks of sourcing information, the critical clinical and empathetic decisions remain firmly in human hands.

"I think AI can make us more human. AI is taking the mundane task of having to source information from different places... giving that time back so that we can be on the phone with you, having empathy and sympathy for your problem, understanding and hearing you."

— Gaston Concilio

Why the Tech-Native Advantage is Crucial for AI Success

Not all PBMs are created equal when it comes to adopting AI. Many industry giants are held back by “legacy inertia”: trying to run modern AI tools on top of infrastructure built in the 80s and 90s that is not structured for modern uses. This is where the competitive landscape is being redrawn; a modern tech stack is required to compute the millions of scenarios involving drug pricing, rebates, and sourcing. 

Adoption of AI shouldn’t just mean using modern technology either; it should be done responsibly and ethically. This responsible approach is baked into the SmithRx DNA. It’s remarkably easy to "slap AI" onto a problem and hope for the best, but we’ve chosen a path of intentional, measured deployment. By conducting rigorous internal experiments, we’ve proven that these tools don't just speed up processes, they improve them. As we expand these tools to our members and partners, we maintain that same level of rigor, ensuring that every piece of information provided is not just fast, but verified, thorough, and genuinely valuable.

“You can't put a Tesla battery into a horse carriage… and have it work. It's important to prepare your infrastructure. That starts at the top of the organization, not just the physical infrastructure, but also the culture of building modern technology and using the available technology to derive business outcomes.” 

— Gaston Concilio

For more on artificial intelligence and its impact on pharmacy benefits, watch the complete Full Disclosure episode, “Beyond the Buzz: How AI Is Redefining PBM Success at Scale”, here. If you’re curious about red, and green, flags to spot vendors’ AI-readiness, download our PBM AI Innovation Checklist.

Interested in how a tech-native PBM can lower your costs and improve member experiences? Connect with our team to see how you can be at the forefront of the pharmacy benefit revolution.