You call a pharmacy because you need a prescription refill. Your mother’s blood pressure medication. You’ve been waiting on hold for 15 minutes. Finally, a voice answers — but it’s not a pharmacist. It’s an AI. And it can’t understand why you’re upset.
This isn’t a hypothetical. It’s exactly what happened to hundreds of customers of Kinney Drugs, a Vermont pharmacy chain that recently pulled its AI phone assistant after a flood of complaints. The company thought it was saving money. Customers thought they were being dismissed.
When a machine becomes the gatekeeper of your health, it’s not efficiency — it’s betrayal. And that’s the real story here. Not a technical glitch. Not a poorly trained model. A fundamental misalignment between what the business wanted (cost reduction) and what the customer needed (trust, empathy, human nuance).
Let me name the company: Kinney Drugs. They’re a real pharmacy with real patients. People who rely on them for medication, for advice, for a lifeline. And when those patients called, they got a machine that couldn’t handle the complexity of a simple question like, ‘I need to talk to a human.’ The AI was designed to handle routine tasks, but healthcare is never routine. Every call carries a story — a scared caregiver, a confused senior, a parent managing a child’s chronic illness.
Here’s what most AI cheerleaders won’t tell you: ‘good enough’ AI in high-stakes contexts is worse than no AI at all. Because when you automate a conversation that matters, you don’t just save money — you erode trust. And trust, especially in healthcare, is the only thing that keeps a customer coming back. Kinney Drugs learned this the hard way. They pulled the system. Human staff returned. Customers breathed a sigh of relief.
But the damage is done. Those hundreds of complaints are now public. The lesson is written in red ink: You cannot automate empathy. You cannot algorithmicize trust. And any business that tries will be punished by the very people it claims to serve.
I’ve seen this pattern before. Companies rush to deploy AI in customer-facing roles, driven by a spreadsheet that shows minutes saved per call. They forget that every minute saved is a human moment lost. They forget that the person on the other end of the line is not a data point — they’re a patient, a mother, a father, a scared human being.
So here’s my position, and I’m not backing down: If your AI can’t handle the emotional weight of a healthcare conversation, don’t deploy it. Period. The risk of backlash, the cost of rebuilding trust, is far greater than the short-term savings. Kinney Drugs is proof. They had to pull the plug. The question is: will your company learn from their mistake, or will you be the next cautionary tale?
The twist in this story is that the technology itself wasn’t broken. The strategy was. The business prioritized efficiency over understanding. And when you strip away the human element from a service that is fundamentally human, you don’t get a better system — you get a crisis. Stop treating customers like tickets. Start treating them like people. Your AI can wait.
FAQ
Q: Was the AI actually broken, or was it just poorly implemented?
A: The AI likely worked as designed for routine tasks, but the problem was strategic: it was deployed in a context where emotional nuance and trust are critical. The technology wasn't the root cause; the decision to prioritize efficiency over empathy was.
Q: What can other businesses learn from this?
A: Don't automate customer-facing roles in high-stakes industries unless you're willing to handle the backlash. Test your AI with real humans in emotionally charged scenarios. If it can't handle a scared caller, it's not ready. And always have a human fallback.
Q: Isn't this just a case of bad AI that will improve over time?
A: Partially, but the deeper issue is that some interactions are inherently human. Even perfect AI can't replace the feeling of being heard by a real person. The improvement needs to be in business strategy, not just algorithms.