Building on Feedback!

Imagine a product or a system, which connects as many areas as possible and has as many features as possible. It seems as if everything has been cramped into one design. Doesn’t do quite well, Does it? Well this was our idea about a conversational agent which detects people’s expressions, their feelings and makes a conversation with them, sympathises with them and encourages users to go to events which would help them grow holistically. Also user could book the tickets through the conversational agent. Main idea was to suggest events to users. We tried to put in as many things as possible in one conversational agent. And during the process for every problem we faced, we added another feature and that was where we had gone wrong!

After the feedback we realised the actual meaning of an actual conversational agent and a prototype. Our product need not have millions of functions, but it needs to have one main function which defines it and all the other features should be supporting features and not features which take product to a whole new direction. In such a case, graph of useful functions becomes quite haphazard. Hence we focused on our main concept- suggesting events to users and encouraging them to go to events. We designed functions and features around this idea to make the product more user like-able, accessible, and to limit the conversation. Sometimes limiting the scope works better for a product and product can be defined in a much better way.

Coding also helped to cut down the features of product making it more specific and a working prototype helped us define our idea in much better way! Coding the product makes a person think critically about all the features as now the features are being converted into action. Since our product was a conversational agent, we had to look up for every kind of possibility to make it feel more like an intelligent bot, otherwise if bot does not have answers for questions, it makes either itself look stupid or user feel stupid for not understanding the product well. We mapped out the complete conversation considering all the directions that conversation can move into. By doing this we had to keep ourselves in the place of user as well as in place of bot. Through this we could define the user journey through conversation in a better way. After mapping of conversation, came mapping of coding. Same sentence cannot be displayed to the same user again and again neither can the same events be shown. Hence we had to make sure that there are no overlaps which ensured a better user experience and coverage of range of conversation.

Overall this week was more about working on feedback to define the conversational agent in a better way! A hazy idea was solidified into a working prototype.