Is Your Chatbot Artificially Intelligent?
The chatter at SXSW this year was all about the chatbot. Several sessions and a lot of side conversations were aimed at helping companies understand what chatbots are and how to use them. I got the sense that everyone wants to have one of these in their product right away.
But while everyone talked about the ways in which these are going to revolutionize user engagement, it sounded too much like “me too!” for my comfort. Too often it appeared to me that companies want to be trendy rather than think critically about whether their product needs a chatbot.
One thing I left certain of is that chatbots are hard to do well. I spoke with a chatbot expert and marketer who cares about brand consistency. They helped me understand more about defending against trolls, and emphasized how hard bots are to execute even before touching on natural language processing!
A Case in Point
Five minutes with the chatbot built into the SXSW app proves how hard a technology this is to master. Abby, the SXSW bot, is a chat interface that brings you to the other areas of navigation in the app but doesn’t add any value on top of that. Think of her like chat-based hyperlinks.
So why are chatbots the next big thing? They are an external manifestation of big data capabilities. As Dave McClure of 500 Startups mentioned in his session: “Data science is being recast as AI. And it’s not there yet.” A lot of talks threw around the terms data science, big data, AI, natural language processing, neural networks, and chatbots as if they are all one and the same thing. But while these things are all related, they are not identical. Want to prove that your company can do the latest and greatest in machine learning? Create a user-facing experience with a chatbot.
Does a Chatbot add Value?
While chatbots can be a manifestation of your machine learning prowess (or a demonstration of a lack thereof), remember the very basic question on any product manager’s mind: will this new feature add value to my product? A poorly executed chatbot (#abby) will provide more pain and frustration to your consumer than no chatbot. Remember the last time you had to call an airline to change your flight and got stuck in a voice recognition circle of hell?
“Please let me know if you want to change your reservation, book a new flight, or look up your flight status.”
“I’m sorry, I didn’t understand you. Please let me know if..”
“I’m sorry, I didn’t understand you. Please let me know…”
“CHANGE THE DAMN RESERVATION!”
A negative chatbot experience leads to a desire to throw your phone out the window. By all means jump into the chatbot fray if it will enhance your product, but remember the age-old adage: friends don’t let friends create poorly-designed bots.