Kinney Drugs AI Phone Assistant Rollback

Kinney Drugs scales back AI assistant following critical failures

Kinney Drugs is reverting its incoming patient call system to a traditional touch-tone model after its AI assistant, "Burt," generated hundreds of customer complaints. The pharmacy chain reported that the AI assistant failed in critical areas, including providing wrong medication dosages, delivering incoherent responses, and missing prescription notifications.

While the company will maintain Burt for outbound communications—such as prescription refill texts—these services now require an explicit opt-in from patients.

Technical and operational failures of the "Burt" AI

Introduced in May 2026, Burt was designed to handle patient communications regarding refills and prescriptions. However, the implementation resulted in significant reliability issues that compromised patient safety and user experience.

Critical errors and user experience

Customers reported that the AI was unable to handle the complexities of pharmacy interactions, leading to:

• Medical inaccuracies: The bot provided incorrect dosage information to patients.
• Communication breakdowns: Calls were described as incoherent, making it difficult for patients to receive necessary health information.
• Notification failures: The system missed critical prescription notifications.

Privacy and compliance claims

In response to customer concerns regarding the exposure of personal health information, Kinney Drugs President John Marraffa stated that the system is fully HIPAA compliant. Marraffa further claimed that the AI is not open-source and "does not generate or manipulate data," though this claim has been met with skepticism by technical observers who note that generative AI, by definition, generates output.

Industry perspectives on Voice AI in healthcare

The failure of the Burt assistant highlights a broader tension between the push for AI automation and the necessity of domain expertise in high-stakes environments like pharmacy.

The role of domain expertise

Industry insiders suggest that while the technology for AI agents exists, the primary bottleneck is implementation and domain-specific knowledge. One pharmacy AI vendor noted that successful deployments require hiring pharmacists as project managers to ensure the AI understands the critical nuances of the industry.

High-risk use cases

Analysts and engineers argue that pharmacy phone lines are an inherently poor first use case for voice AI due to several factors:

• User demographics: A high volume of older callers who may have less patience for AI interfaces.
• Complexity of data: The difficulty of accurately processing complex drug names and insurance details.
• Zero tolerance for error: Unlike general customer service, errors in medication dosage can have life-threatening consequences.

Synthesis of community critique

Discussion among technical professionals suggests that the Kinney Drugs incident is part of a larger trend of "shortsighted decisions" by leadership prioritizing cost-cutting over customer experience.

"Companies will pawn off their customers to AI at their peril. In many ways this is a repeat of the India call center train wrecks of the 00s... AI is just that story of shortsighted decisions by weak leadership playing out all over again."

Other critics argue that for regional pharmacies, personal human service is a primary competitive advantage that is eroded when replaced by automation. There is also a prevailing sentiment that many companies are deploying "hacky" LLM implementations that are not yet capable of delivering repeatable, specific outcomes in production environments.