Kinney Drugs отключила ИИ-ассистента на телефонных линиях после сотен жалоб клиентов

Kinney Drugs pharmacy chain disconnected its AI voice assistant from phone lines on August 7, 2026, after receiving hundreds of customer complaints, according to TV channel WCAX. The company has returned to its traditional incoming call handling scheme.

What went wrong with the AI operator

Kinney Drugs' AI assistant failed to handle real customer requests — the chain received hundreds of complaints and decided to quickly shut down the system. Kinney Drugs is a pharmacy chain in the northeastern United States, operating in the states of New York and Vermont. The company had implemented a voice AI assistant to handle incoming calls — one of the busiest channels in the pharmacy business.

The chain received hundreds of complaints about the AI assistant's performance and decided to disconnect it — according to a WCAX report from

August 7, 2026.

The company has not published official statements about the nature of the complaints. Typical problems for voice AI operators in retail include: the system fails to recognize non-standard phrasing, cannot verify a customer against the prescription database, freezes on complex questions, or fails to transfer the call to a live specialist. In a pharmacy context, such a failure is especially sensitive: a customer who could not get through or received incorrect information about a medication may make the wrong medical decision.

Notably, the company responded fairly quickly — which suggests the complaints were not isolated and came through multiple channels simultaneously.

Why a pharmacy is harder than regular retail for an AI operator

Voice AI systems handle standard call center requests well — business hours, addresses, delivery status. The pharmacy context adds layers of complexity that developers often underestimate:

• Customers often do not know the exact trade name of a drug or confuse a brand with an INN (International Nonproprietary Name)
• Inquiries about prescription drugs require customer identification and access to their history
• Incomplete or incorrect information directly affects health outcomes
• Elderly patients — the primary audience of pharmacies — often struggle to adapt to voice interfaces
• Many calls are urgent in nature: the medication is out of stock and needed today

US pharmacy chains are actively deploying AI operators as a way to reduce staff workload and cut costs. Kinney Drugs is not the first — and most likely not the last — company to discover that saving on live operators costs more if the customer experience deteriorates sharply.

A critically important pattern that is violated in such implementations: the customer must have a clear and quick path to a live operator at any point during a conversation. When this is absent, frustration accumulates and complaints go not to the AI system but directly to the media.

Why the case attracted developers' attention

News about a regional pharmacy chain from Vermont gathered 114 upvotes and more than 126 comments on Hacker News — an atypical result for a local TV report. This indicates the story struck a nerve: developers and product managers see it as a case study on the limits of automation in sensitive services.

Discussions around such cases invariably surface the same themes: the hidden costs of poor customer experience, the problem of insufficient escalation to a live operator, and the gap between "technically functional" and "customer satisfied." Automation that saves money on the backend but degrades the experience of a patient — especially a vulnerable one — ultimately costs the company more.

What this means

The Kinney Drugs case confirms: voice AI assistants in pharmacy chains require more thorough testing and phased deployment than in most other consumer services. When the core customer base consists of elderly people with prescription needs, the acceptable failure rate for an AI operator is significantly lower — and accordingly, so is customer patience. Quickly shutting down the system after hundreds of complaints is the right response, but the better outcome would have been to prevent the situation in which those complaints accumulated in the first place.