Formulary management is one of the most consequential functions in healthcare operations. The decisions made by Pharmacy and Therapeutics (P&T) committees about which drugs to cover, at what tier, and under what conditions directly affect patient access, plan costs, and rebate revenue. Yet the tools supporting these decisions have not kept pace with the complexity of the task.
For most health plans and PBMs, formulary management still looks remarkably similar to how it looked a decade ago: analysts pull data from multiple systems, manually cross-reference drug classifications, review pricing updates in spreadsheets, and compile recommendations in Word documents for committee review. A single formulary change can touch hundreds of related drugs, interact with dozens of rebate contracts, and affect thousands of members. Doing this analysis by hand is slow, error-prone, and fundamentally unscalable.
Where AI Enters the Picture
AI does not replace the clinical judgment of a P&T committee. That is a critical distinction. What AI does is compress the analytical work that precedes and supports those decisions. Instead of an analyst spending two days compiling the impact analysis for a proposed tier change, an AI-powered system can generate that analysis in minutes, pulling from drug pricing data, utilization history, rebate terms, and clinical guidelines simultaneously.
The key areas where AI is already proving its value in formulary management include:
- Automated rule interpretation. Health plans maintain hundreds of pages of formulary rules covering prior authorization criteria, step therapy requirements, quantity limits, and exception protocols. Natural language processing can parse these documents, extract the conditional logic, and codify it into executable rules that a system can evaluate automatically.
- Impact simulation. Before making a formulary change, plans need to understand the downstream effects. AI models can simulate the financial, clinical, and membership impact of a proposed change by analyzing historical utilization patterns, rebate structures, and member behavior data.
- Drug classification and mapping. New drugs enter the market continuously. AI systems can automatically classify new NDCs, map them to existing therapeutic categories using data from sources like Medi-Span, First Databank, or Micromedex, and flag potential formulary placement recommendations based on the plan's existing rules.
- Regulatory compliance monitoring. CMS updates Part D coverage requirements annually. AI can continuously monitor regulatory changes and flag formulary configurations that may fall out of compliance, rather than relying on manual review during the annual submission cycle.
The Data Foundation
AI is only as good as the data it can access. The healthcare industry is fortunate to have a growing ecosystem of high-quality, often free, pharmaceutical data sources. The openFDA API provides access to drug labels, adverse event reports, and recall data. RxNorm offers standardized drug naming and relationship data. DailyMed publishes current prescribing information for virtually every marketed drug. Commercial data feeds from Medi-Span and First Databank provide pricing, GPI codes, and therapeutic classifications.
The challenge is not data availability. It is data integration. Most health plans and PBMs have these data sources scattered across different systems, updated on different schedules, and accessed by different teams. An effective AI-powered formulary system needs a unified data layer that normalizes and connects these sources into a single queryable model.
From Filing Cabinet to Intelligent System
The evolution of formulary management technology can be described in three phases. The first was paper and spreadsheets, where institutional knowledge lived in binders and the heads of experienced analysts. The second, which most organizations occupy today, is database-driven: tools like MMIT Navigator, RxBenefits, or custom-built systems that store formulary data in structured formats but still require manual analysis for complex decisions.
The third phase is AI-native, where the system does not just store information but actively supports decisions. It ingests new data automatically, flags anomalies, suggests optimizations, models outcomes, and surfaces the information that a decision-maker needs at the moment they need it. This is not a futuristic vision. The underlying technology exists today. The gap is in implementation, and specifically in building systems that understand the domain well enough to be trusted with these decisions.
What This Means for Health Plans
Organizations that adopt AI-powered formulary management gain three specific advantages. First, speed: decisions that previously required weeks of analysis can be supported in hours. Second, consistency: automated rules application eliminates the variation that comes from different analysts interpreting the same policy differently. Third, scale: a single system can manage hundreds of formulary variations across different plan designs without proportional increases in analyst headcount.
The organizations that will benefit most are mid-market health plans and PBMs that manage enough formulary complexity to need sophisticated tools but have not historically had the budget for large custom development projects. AI is lowering the cost of building these systems, making enterprise-grade formulary management accessible to a broader market.
The shift is underway. The question for health plan leadership is not whether to adopt AI for formulary management, but how quickly they can move from evaluation to implementation.