Chatbots are all the rage these days, becoming the latest industry buzzword after “Cloud” and “Big Data”.
I am no stranger to chatbots. In fact, Originally US recently built a chatbot for Mediacorp – Love Preparatory 预习恋爱 as a tie in to their latest blockbuster, Dream Coder, on top of its role as the technical consultant for the show.
I’ve also experimented with chatbots in the past, believing then that having a natural language interface for users to interact with their computers is the future. To illustrate that concept, I made this in 2003; when I was just 15.
Ultimately, a chatbot is just another user-interface. There are plenty of cases where using a chatbot would positively impact end-users, and there are cases where we believe people are building chatbots for the sake of building one.
Let’s take a look at the recent chatbot sensation, Bus Uncle.
We found that it took more than 30 seconds for a user to locate the bus arrival timing of a specific bus at a specific stop – Service 48 at Bedok North Depot. It takes even longer if you factor in the time taken to launch the Facebook Messenger app and typing in “bus uncle” to locate the bot before initiating the chat.
Comparatively, the same interaction on our SG BusLeh app took way less than 15 seconds.
Better yet, if you have already added the same bus-stop to favourites, it takes just a single tap and barely 2 seconds to get the latest arrival timing for service 48. This is because SG BusLeh automatically refreshes all bus timing upon app launch.
Bus Uncle can also implement a “Favourite stops” mechanism to emulate the efficiency of a well-designed visual interface, but the question remains, do we need a chatbot interface to achieve that level of interaction efficiency?
At the end of the day, the rules in determining when to use a chatbot are simple:
A majority of your users are first-time users
First-time users are not familiar with your system yet, so having a chatbot to guide the users along helps them access the information or transactions they need.
A majority of your users only make the same transaction or query once
If you are going to conduct the same query or transaction as much as multiple times a day, it is much faster for you to tap and scroll on a visual interface that you are familiar with rather than saying the same things to the bot many times over.
Traditional interfaces cannot sufficiently convey what’s behind the mystery meat
When you need substantial explanation to help users understand what is behind each each choice or options, a chatbot can fulfill that task without confusing the users.
When you have a non-linear choice/navigational/information hierarchy, and human language is the most natural way to help users cut through the hierarchy layers
Visual interfaces excels in presenting linear information hierarchy, but once you have a confusing and massive decision tree, having a chatbot that understands what your users want help cuts through that mess.
When it is faster to chat or type than swipe or tap.
Interested to build a chatbot for your business and not sure if it is the most effective approach? Contact us for a free consultation today!