A recent post on the TechCommGeekMom blog contained a link to an article titled “Chatbots are not the future of Technical Communication” (everypageispageone.com/2018/01/30/chatbots-are-not-the-future-of-technical-communication/), written by Mark Baker.
TechCommGeekMom provided her own summary of the post, saying that she “didn’t think chatbots were really that much of a tech comm thing either”. I was intrigued and decided to read the article for myself. If I had found Mark’s blog everypageispageone.com when I was looking for blogs to follow, I would have added it to my list. I am following it now!
A ‘chatbot’ is a computer program designed to simulate conversation with human users. In its simplest form, it will scan for keywords in the user input (text or auditory), then pull a reply with the most keyword matches. I would strongly suspect that the calls I get from perky “Heather” at “Account services” (company unknown) are more than likely some form of a chatbot program.
Mark lays out a thoughtful argument that even though “every tech comm and content strategy conference seems to be about getting your content ready for chatbots”, they are not the future of technical communication. He points out that chatbots may be great at ordering something online but they are not even close to being able to help with troubleshooting a technical problem. Chatbots are not smart in the sense that they don’t have common sense, they are simply a command line interface (CLI). All they can do is respond to the user’s input, and sometimes the responses they give are wrong. It cannot listen to the user’s explanation and suggest alternate solutions. A chatbot cannot show the user anything, and it cannot see what the user is doing. The user has to tell the chatbot what they are doing, but the user cannot tell a chatbot what the problem is because the user is only seeing the result of the problem. Making the chatbot smarter will not solve this.
Mark argues that text is the most meaningful way to deliver content to the user. Text is the type of content most amenable to the various ways in which users search for information. As Mark says “Nothing else lets you speed up and slow down, go straight or turn left with anything like the same ease”. He includes ancillary media – maps, graphics, animations – as playing a valuable role, but points out that it is ultimately text that leads to these. Text allows the insertion of hyperlinks, which can instantly expand the search and allow the user to follow leads. Interaction with a chatbot will limit the user’s ability to change pace or direction at will when looking for information.
I found it sobering to consider that the hope of some of the current boosters of AI is that the technology can be developed to the point of “delivering the right content in the right format to the right person at the right time”: a Nurnberg funnel*. If it was even possible that technology could be developed to that level, it would mean that the user would no longer have to search for information or learn new tasks. But I believe it is essential that users be allowed to seek information in the way best suited to them. Every individual has their own style of learning.
I agree with Mark and TechCommGeekMom (Danielle M. Villegas) that chatbots are not likely to replace the traditional sources for technical communication anytime soon. A chatbot can be a useful tool but it cannot replace human interaction, or replace (as Mark states) the form of information most suited to creating a navigable, searchable field of information – text.
*A legendary funnel through which knowledge could be poured into the head of the learner. It was said to make people wise very quickly when the right knowledge was poured in