Recently I had an insightful communication with a chat bot. A payment provider wanted me to pay for ordered goods and sent me a reminder via e-mail. Unnecessarily so, because I had already returned the goods. Obviously the payment provider didn´t know about it, so I visited their website, which prominently provided a chat bot.
In order to quickly fix my issue, I chose the offered communication channel, and found it rather automated. Questions about order data appeared in a millisecond — as well as the information about invoice amount and payment terms.
But when I answered, "No, I won’t to pay since I have returned the goods already", the chat bot fell silent. No quick automated answer this time. “I need a moment to think” it stated. To be precise, the moment took 12 minutes.
Finally, “Kristin” answered. Friendly and competently, she informed me that I should have reported the return myself to the payment provider. Good to know. Presumably, it´s said so in the fine print. But who reads that? Kristin provided me with a link to register the return and closed the issue with “a nice day”.
Chat bots can be an exciting option for companies that want to limit involvement of human resources for simple and redundant tasks in order to increase efficiency. However, the example above also shows that the potential by far isn´t leveraged yet. On the other hand AI isn´t just about adding a few more if/then algorithms.
So far, though, I cherished the human touch above efficiency. But the chat with Kristin was friendly and efficient. "Thank you. You are a great bot", I resumed. "I am not a chat bot. I am human!” Kristin replied with a smiley face.
In a recent Hackathon, our developers succeeded leveraging AI to personalize our product recommendation feature based on machine learning and Microsoft Azure technology.