Have you ever Been Stuck in AI-powered Customer Service Hell?

The other day, I spent over 20 minutes canceling a fully refundable airline ticket, 19 minutes of it trying to get around the site’s “chatbot” and get to a human!

The task could have easily been handled on the website by clicking a button or selecting an item from a drop-down menu. But like many businesses these days, this one — which shall remain nameless — felt compelled to use a chatbot, to be “innovative”.

Unfortunately, chatbots do little good for most of us who know how to use “Frequently Asked Questions” and/or get the current balance of our accounts online. We call because we have a genuine problem that is too complex to explain to a chatbot, often requiring an action that the chatbot is not authorized to perform, probably because the business doesn’t trust the AI.

And there are a lot of us who don’t want to deal with Chatbots. According to a recent survey conducted by TCN seven out of ten customers prefer talking to an agent when dealing with customer service. We tolerate the wait times, the sappy music, and folks answering from four corners of the world with strange accents, as long as we can eventually talk to a real person.

And yet, a recent investigation by the wall street journal indicates a rush to replace customer service agents with machines by firms large and small, presumably to save money, or improve employee satisfaction. In fact, they end up saving nothing — since eventually a human being is needed to respond to the call — and risk losing customers and frustrating employees who must follow a machine-created script that may not be appropriate for the situation they are dealing with.

Worse yet is when the user is forced to use a chatbot to handle mundane tasks that can be completed via the push of a button: canceling a subscription, suspending a delivery, or changing a reservation. A lengthy battle with a chatbot to do something that can be done by clicking a few buttons is like using a jackhammer to hang a picture on a wall.

Like any other productivity tool, the so-called Artificial Intelligence (AI) — really Machine learning — can improve efficiency and productivity. But indiscriminate use of it just to demonstrate forward-thinking does nothing but frustrate the customers and the employees. The practice reminds me of the time when everyone wanted to “computerize” their business, regardless of the utility and/or cost. Eventually, most businesses accepted the wisdom of the industry consultants that for computers to improve the bottom line or customer experience, their processes may have to change. That change of course necessitated hiring consultants to advise firms on automation. And a new business was born: IT consulting.

What we need is old fashion commonsense. We need to ask ourselves one single question: Does adding AI, add value? If the answer is no, then we need to move on. Technology for the sake of technology doesn’t make sense.

User expectation is a key aspect of the long-term success of any new technology. When they fall short of people’s expectations, they are discarded. In the case of chatbots, despite their deceptively polished outputs such as the impressive writings of ChatGPT, they are good for only a limited number of applications. Pushing them too far will break them and ultimately ends up with people rejecting them — Alexa anyone?

As for me, mercifully, Amazon still allows me to cancel an order with the touch of a button, Hallelujah!