Conversational Commerce : Journey from understanding users to suggesting them.
What comes to your mind When You hear the word bot?
The techie like me would surely think of the Artificial Intelligent system, that can understand the way humans speak (Not sure about others). We started making our bot with this image of our end bot. We worked 24/7, read research papers, online tutorials and did what not to make this a reality. And yes, we ended up creating something quite smart, answering all our questions.
Next, it was time for our bot to interact with its first user and all was shattered. Bot answered something completely random. Then Followed a series of events where one after the other it failed to prove its worth. We wondered everything seem to be working perfect but what happened? We came up with following hypotheses:
- Since, the end-users were unaware of bots. They spoke like it was just another human answering. Some part of the problem was at our end, which Could be fixed.
- Facebook bots usage results pointed OUT A 90% drop rate of at first step itself, if users are asked to type something.
- Further, we followed an approach where we were trying to understand the user intent and respond accordingly. This is creating something like the Google Now, Siri, Cortana is the smaller scale. When it is argued AI is not ready for consumer facing things.
As per our understanding and analysis, all the above factors collectively led to disappointing results.
A new dimension in bot building
We took few steps back and tried going deeper in understanding the purpose of the bots. We moved to the idea of an assistant that very well understands his master and gets his tasks done in most seamless way possible. We changed to the philosophy of understanding the user from previous actions and showing him the relevant information and actionable to get his work done. With this approach we wanted to give users a sense of control, where he can drive the conversation and at the same time we tried to restrict him under limitations of the chatbots.
At current stage this is again hypothesis that this newer version would be more useful and delightful in terms of experience. The early results seem positive but more is needed to come to any conclusion.
Finally I wish to conclude by saying, since there are no set of rules written on the UX design of bots, bot builders have to try multiple ideas and see whatever works best. We forgot that humans are familiar with the language of clicks and taps. So, instead of making bot learn something human we should have used a common medium for both and focused on building a personal connection between a bot and master. I firmly believe more than the technology, the experience matters for the end-user.
PS: I would soon be publishing the next part of this, where would love to share the results of the new version of our chatbot very soon.
MindIQ.in brings you the quickest solution to on-board and run your business on any conversational platform. Interested in building your business over chat?
Sign-up for a demo now at www.mindiq.in
