If you are interested in reaching customers that you can’t get to now, in offering user experiences that go beyond the rigid orchestration of the point and click paradigm, in providing just the right data that someone needs, it may be time to consider creating a chatbot.
When I ran across the description of Colin Megill’s course for O’Reilly Media about how to build your own chatbot (which is starting on July 13), I thought of Eliza, one of the pioneering chatbots that did a serviceable job of parsing sentences and tossing them back to you. Eliza made it seem like you were having a conversation if you got lucky. If you said something it couldn’t handle, the conversation quickly got dull or non-sensical.
Not so with the best of the new generation of chatbots. The world has grown up. Services like Alexa and Siri are leveraging advances in machine learning to improve upon previous user experiences, and, more importantly, such platforms available to developers via APIs. Companies like Facebook and Amazon can offer you a rich context about a user preferences, history, and behavior.
Ask Amazon’s Alexa to order more detergent and it can tell you, “You ordered Super Clean detergent eight weeks ago. I can get you another box in two days for $25."
But more than that, Facebook Messenger, the Apple and Amazon platforms are available to hundreds of millions of people. If you have a good app you want to be there. "If you are a brand, you want your product to be present in that environment,” said Megill.
But this is just the beginning. NPR is building an chatbot called Carebot that will help journalists navigate statistics about articles published online. Megill sees a particularly bright future for chatbots in business intelligence. "Anyone interested in data driven decisions can learn from the Carebot,” said Megill. "It is about getting people involved in data and bringing it to them in a way the can use it."
Of course, as with any hot area, there are a whole lot of chatbots being released that don’t do much. “We are seeing a lot of really trivial bots out there,” said Megill, CEO and one of the founders of pol.is, which is building a Slack bot to increase the enterprise reach of its platform for large scale conversations. “Like any new medium people are just wanting to get something out the door. It is the same thing we are seeing with Virtual Reality."
"Bots that attempt to move beyond trivial interfaces and rigid menus will find that extraordinary combinatorial complexity immediately emerges," Megill said. He believes startups like x.ai are able to handle the complexity of natural language processing (NLP) using machine learning. But there are plenty of worthy bots, like Carebot, that deliver a useful experience without getting too complex.
Megill’s course is designed to teach you to build chatbots that hit the sweet spot between too simple and too complex. Eliza was too simple to matter. x.ai may be too complex for most of us to build. But the chatbot you create will be just right, and may open a new world of customers to you.
Follow Dan Woods on Twitter:
Dan Woods is on a mission to help people find the technology they need to succeed. Users of technology should visit CITO Research, a publication where early adopters find technology that matters. Vendors should visit Evolved Media for advice about how to find the right buyers. See list of Dan's clients on this page.
