How an AI Chatbot Pushed a Client to Competitors

Last week, I was trying to resolve an issue with a bank. A simple task that used to take three minutes of chatting with a live person. The chatbot ran me in circles for twenty minutes, not understanding the essence of the problem. It stubbornly didn't want to connect me with a person, and in the end, gave me a link to the FAQ, where there was no answer either. I irritably closed the app and thought it was time to change banks. Although I've been banking there for about eight years…

\ Have you noticed how people have started reacting to innovations in the form of chatbots, which supposedly should bring benefits not only to companies but also to clients… Companies are massively implementing AI bots in customer support, and more and more often I observe the same reaction in people. Irritation. Rejection. I perfectly understand business logic. Reduce support costs, optimize operations, scale without hiring, and so on. On paper, it looks perfect. A bot processes thousands of requests simultaneously, doesn't take a vacation, doesn't get sick, and works around the clock. Management sees the numbers and says yes, this works, let's continue!

\ But here's what actually happens… Chatbots only understand basic requests from a pre-written list. Any deviation from the script and the circus begins. "Sorry, I didn't understand your question.. Try rephrasing." The person rephrases. The bot doesn't understand again. Suggests going to the help section. In the help section, there's a bunch of information in the "question-answer" style, and you want to get an answer to your question, but they make you go through 6 circles of hell. And let's be honest, most of these bots are frankly stupid. When a person contacts chat with a problem, they're most likely already dissatisfied with something, irritated, and this "robot employee" can make an already bad situation worse.

\ There are areas where bots really work normally. Order tracking, balance checking, and simple FAQs like "what are your business hours". Standard informational requests without emotional load. Here, automation makes sense. But when a person has a problem, when something goes wrong, when they're upset or angry, a bot is a disaster. Because in such moments people need not information but understanding that their problem is heard and they're trying to help them solve it.

\ It's worth adding that some companies try to "mix" bots and real employees without warning the client who exactly is communicating with them, and most often starting communication with a bot under a real person's name and then (possibly) a real "live" person actually connects. This approach raises even more doubts because often it's terribly implemented and you think either a chatbot is communicating with you the whole time, continuing to write "connect me with a person," or you're already communicating with a person who responds to you like a chatbot, and besides irritation as in the first case, now people get the feeling they want to deceive them (and actually that's true).

\ I conducted analysis on several projects where we looked at long-term loyalty metrics. And here's what's interesting. Companies that aggressively automate support save money in the first six months. Then, quiet customer churn begins. Not massive, not obvious. This often depends on how much support people need on a particular issue and whether they can solve it in 1-2 messages or need a real, live, several-minute dialogue with a person. As a result, people stop renewing subscriptions. Don't recommend to friends. When an alternative appears, they leave without hesitation. This may sound somewhat "poetic," but it's true: the emotional connection is broken. The brand turned into a faceless system. And a faceless system isn't a pity to abandon. After all, we act for the most part based on our emotions much more often than on reasoning, "~~this bank has 0.1% lower commission, so I'll stay with them even though their chatbot kills my nerves with its stupidity and can't solve my problems,~~" rather, everything is exactly the opposite.

\ It's funny that we spend millions on building a brand, on creating an emotional connection with the audience, and on marketing that should show we care. And then we put a bot whose very existence broadcasts "you're not important enough to us to hire a person." This is the last link in the chain. A person went through your entire marketing funnel. And now they need help. This is the moment of truth. The moment when a real attitude toward the brand is formed. And if at this moment they're met by a "soulless bot" (and most often just stupid) that can't solve a simple task, all previous investments in marketing go down the drain.

\ I recently wrote about how AI kills conversion in advertising with its machine presentation of creatives in targeting. People learned to detect robots and ignore them. It's like when you're driving down the street and there are hundreds and thousands of advertising banners around you, but you don't pay attention to any of them. Same story here, only worse. Because in advertising, you lose a potential client. And in support, you lose an existing one. Who paid money and could have paid for years more? It's actually interesting to observe how the pendulum starts swinging in the opposite direction. This totally digital world, where everything is automated and optimized, is starting to bore people. I see this by various signs.

\ This doesn't mean you need to abandon technology. The question is balance. A bot can be the first line for simple questions. But the transition to a person should be instant and obvious. Not through three menu levels, not after filling out a form, but right now. And if a person immediately wants communication with a person, allow this to happen. Don't torture them with these "~~AI rituals~~." Because when a person needs help, every extra minute works against you. And a bot that runs you in circles isn't just an inconvenience. It's a message. "We don't want to spend time on you"

\

\ It's funny how it turns out. We live in an era when technology allows us to be closer to clients than ever. Know their preferences, anticipate needs, personalize experience. But instead, we use these technologies to distance ourselves.

\ In the long term, this is always a losing strategy. Because clients aren't just "numbers in your CRM" - they're real people who want to deal with people. Especially when something goes wrong. You can save on support today. But tomorrow you'll lose clients who could have brought money for years. You'll lose word of mouth. You'll lose loyalty that can't be bought with any marketing. And the most offensive thing is that this loyalty isn't killed by a bad product. It's killed by a chatbot that couldn't answer a simple question.

\ ==What do you think about this? Interested to know your opinion. Thanks for your attention.==