When an AI server says “I’m here to help!"—they aren’t here to help you—it’s there to tire you out until you accept your fate or log off in frustration.
Companies today, in their never-ending quest to cut costs, have increasingly turned to AI-powered solutions for customer service—so they (i) hire fewer human agents, and (ii) make it frustratingly difficult for you to reach an actual person.
Back in the old days, AI customer service was nothing more than keyword-triggered auto-responses and simplistic, rule-based bots. Today, however, these systems—powered by sophisticated large language models—have become annoyingly competent at handling basic inquiries. So now, companies aren’t just deploying AI to streamline services; they’re intentionally adding friction to prevent customers from easily reaching human agents.
Not that they haven’t done so. It’s just that it has become significantly worse and annoying.
Most, if not all customers, conditioned by decades of bad chatbots, naturally assume AI is going to be useless. So now people immediately type something like, “I want to talk to a human!"—repeatedly, with increasing desperation. Companies, anticipating exactly this scenario, conveniently raise the bar for human intervention. Previously, a simple request like “I need to speak to a person” was sufficient to escalate your case. Now, AI gatekeepers are programmed with elaborate conditions: “Transfer to human agent only if the customer is angry, impatient, repeating themselves endlessly, uploading desperate screenshots, or displaying clear signs of emotional breakdown.”
Meanwhile, companies—fully aware that consumers neither trust nor expect much from AI (especially when sensitive issues like money are involved)—have zero incentive to improve AI quality. Instead, they opt for the cheapest, third-party AI solutions that money can barely buy. So now we’re having slow, clunky chatbots, strategically placed to form an impenetrable wall between customers and actual customer satisfaction.