Part 1 of a 4-Part Series: What is AI? An Overview, Its Role in 340B, and How You Should Start Thinking About It

February 19, 2025 Sponsored Content Views

Leveraging AI and Understanding Its Limitations

What is AI?

Artificial Intelligence (AI) has become a buzzword in nearly every industry, but what does it actually mean, and how does it impact healthcare? At CompassRx, we believe that understanding AI is the first step in effectively leveraging it—explaining how it works, where it’s useful, and where it still has limitations.

AI: Fact vs. Fiction

When people think of AI, they often picture movie depictions like HAL from 2001: A Space Odyssey or J.A.R.V.I.S, Tony Stark’s AI companion from Iron Man. In reality, AI isn’t actually intelligent the way humans are. Instead, AI is all about pattern recognition—computer models capable of analyzing massive amounts of data to detect trends and then generate the right response based on what it has learned.

Breaking Down the Basics

AI models function by processing structured (excel, word files) and unstructured (audio, visual files) data to identify patterns and determine appropriate responses. In healthcare, AI models identify patterns using 3 common types of data:

Images (Computer Vision) – Analyzing medical images to help identify small fractures in an X-ray or hard to spot tumors in an MRI
Text (Large Language Models) – Powering tools like ChatGPT which analyze large bodies of text to answer questions or create documents
Audio (Voice Models) – Identifying vocal patterns to convert speech to text for applications like scribing

Some advanced AI systems, known as multimodal models, integrate two or more of these data types to handle more complex tasks. For example:

Audio + Text: AI-powered translators combine voice and language processing to enable real-time medical interpretation between English and Spanish speakers.
Text + Images: AI can synthesize data from images and text, making it possible to extract information from scanned medical documents and respond to questions about them, such as answering a question about an x-ray.

The Biggest AI Misconception

One of the most common misunderstandings about AI is that it can learn and adapt on its own. The truth is, most AI models aren’t self-learning — the more these models are used (and corrected), the better they get, but always with a healthy level of human oversight. That’s why CompassRx uses data from our learnings to carefully fine-tune our models over time, ensuring the covered entities we work with are using the most accurate, safe, and effective models.

Does AI Belong in 340B?

Yes! In fact, AI is already streamlining healthcare operations in several ways, including:

Document / Data Analysis – AI supports your 340B audit readiness by automatically reviewing all of your records (instead of manually auditing a select few) to help you stay compliant
Generating Healthcare Documentation – AI auto-generates Rx prior authorizations to ensure your patients get their medications approved on time
Discrete Task Automation – AI helps with scheduling, 340B referral capture, and other administrative tasks to reduce burden on providers and administrators
Empathetic Communication – AI-powered engagement tools to reach out to patients and help with greater 340B in-house pharmacy capture
Language Translation – AI facilitates communication between providers and patients who speak different languages to break down barriers of care.

So Can AI Really Do It All?

While AI can offer many efficiencies, it also has limitations—especially in clinical decision-making. Think of AI like an overconfident intern who wants to answer every question, even when unsure. They could have the best of intentions and still misdiagnose seasonal allergies for the flu! For example, search engines using AI have been known to provide misleading medical advice, similar to the pitfalls of WebMD but amplified.

To mitigate these risks, AI models need to be trained to recognize their limits. The right approach involves:

Defining Boundaries – Instead of always providing an answer, AI should be programmed to escalate complex medical questions to trained professionals.
Human Oversight – AI should serve as a support tool, not a replacement for expert judgment. In critical areas like patient safety, pharmacists and healthcare teams must remain in control.

The Bottom Line: AI as an Ally, Not a Replacement

When implemented responsibly, AI can be a powerful ally for healthcare teams—boosting efficiency, enhancing patient care, and reducing administrative burdens. However, its success depends on having a well-structured AI strategy that prioritizes safety and human oversight.

Curious about how AI can support your organization? CompassRx offers free AI strategy consultations to help support your covered entity safely and effectively.