The homepage contains the chatbot’s book recommendations to the user. But for first time users, the page just shows placeholders. The chatbot then approaches the user with an invitation to discuss book preferences.
 The conversation takes front and center stage in order for the experience to feel more personal. It would be as if you were talking to your friend about books you’ve read.
 Each user response is parsed for keywords for searching through Google Books API. The chatbot’s response, which comes from a database of questions, is then displayed on the front-end. The current prototype simply follows a linear sequence of questions, independent of user responses. Future development will improve this algorithm with natural language processing and other machine learning algorithms for finding the best response depending on the user’s conversation.
 The conversation ends with a link to the recommendations.
 Book recommendations are displayed as book covers using Google Books API.
 Clicking on each book cover reveals book title, author, description, as well as the Google Books search query that brought that specific book. This helps the user understand why they are getting the recommendations and can inform future conversations.
 If the user accidentally leaves mid-conversation, a message displays instructions on how to get back.
 Responsive layouts were also designed for viewing on mobile devices.
 Inital brainstorming identified key questions to ask users and how the chatbot responses could be generated based on the user’s conversation. Questions were then stored in a database.
 Mobile first storyboarding identified the main sequence of webpages to design. Color palettes and logos were intentionally avoided so that the general layout could be resolved first.
 Crazy 8’s ideation was used to brainstorm logo ideas that exemplified the website’s identity: personalized content, emphasis on conversation, and empathetic guidance. The final logo was chosen for its familiarity with conventional chat interfaces. The skeumorophic design reflects how the chatbot should feel like conversing with a friend.
 A warm color palette was chosen to provide a comforting environment for personal conversation.
prev / next