Chatbots: Creating a Better Experience? It’s a Piece of Cake


Last week, a popular grocery store in the southeastern United States made national news over a cake order.  A proud mother used the online ordering system to request a custom cake. The cake was supposed to read: “Congrats Jacob! Summa Cum Laude class of 2018.” The online tool censored a word on the cake it deemed obscene, replacing it with three hyphens.  As it turns out, the word censored is the preposition “with” in Latin.

As I read about the cake kerfuffle, I felt there were a series of missteps on, not only the grocery store side, but the customer’s side as well.  Missteps that could have been avoided with a conversational chatbot with the option to transfer directly to a live agent.

The word was omitted immediately on the form.  The customer ordering the cake added details regarding the meaning of the phrase in the special instructions section of the order form.  Had this been a chatbot, transferring to a live agent could have provided clarity with a human touch.  Because of the absence of this option, the customer could have called the bakery directly to resolve the issue.

The bakery, wrote verbatim what they saw on the form, assuming that’s what the customer wanted, ignoring the special instructions.  This time, a call from the bakery to the customer to confirm the customer’s wishes for what the cake should actually say could have alleviated the newsworthy situation.

I’m surrounded by an amazing group of NLU and self-service application experts who really know UX/UI design, IVR and ITR application development, computational linguistics and system integrations. I mean, they designed a chatbot that can suggest the right wine when you send it a chicken emoji.  There is no doubt our team of NLU experts are passionate about the potential AI has to empower consumers and easing common tasks — like ordering a cake online without any issue.

I wondered exactly how Aspect’s NLU chatbots would handle a situation like this. More accurately, I wondered if our team of NLU experts had ever considered this exact scenario. I connected with chatbot expert, Tobias Goebel, Senior Director of Emerging Technologies for his thoughts on this situation.

Tobias started by reminding me that language is complex and nuanced. Aspect NLU has linguistic intelligence that includes a vast lexicon of all the words and expressions of a language, across 10+ languages. That’s pretty impressive.

He continued by explaining the parser brings the intelligence needed to go beyond simple word spotting to help with language-related tasks as part of business processes. And while it’s not on par with human intelligence, it would’ve done a better job in this case.

Tobias ended by saying, “when you use technology to help you run your business, make sure to use technology that is actually suited for the job.”

Creating a better experience with chatbots can be a piece of cake!

Learn more at the Aspect NLU Lab.