#5a – Rationale of Medium Fidelity Prototyping Approach

1. Our prototype is more vertical than horizontal. This is because chat and map are crucial in testing our prototype. Since making a dynamic chat and map are time consuming we decided to reduce horizontal features of the prototype.
2. The chat is Wizard-of-Oz’d by having a human that’s pretending to be a bot. The bot’s responses will be scripted so that we’re consistent throughout. From a technical standpoint, we’ve already created a chat and map interface. We used ZenDesk (a customer service company) and Google Maps API. We will be telling the subject that we’re using a Google voice API to parse their chat. We will then reveal that a human is the bot.
3. Functionality is more important than appearance.
4. Prototype should be used on a phone.
5. For our prototype, we are using the Google Maps API and ZenDesk.

#5b – Prototype Demonstrations

Figure 1: The prototype consists of a map screen that covers the entire interface with basic controls. The map is built using Google Maps API, allowing the map to be dragged, zoomed in/out, clicked for details, and switched to Street View. At the bottom left corner is our chat button, which was kept small to minimize it being intrusive in any way as we have found most users tend to mainly focus on the map.

Figure 2: The chat window allows the user to communicate with our bot. This screen shows the startup greeting and appearance of the chat window. The top right button shrinks the window so that the user may return to the map view at any time and continue exploring.

Figure 3: This shows an example of the conversation flow one might expect to see from the user interacting with the bot. Given what information the user requests, the chatbot will respond accordingly and make any necessary updates on the map view and direct the user’s attention to them.

Figure 4: This screen shows the updated map view according to the conversation between the chatbot and the user. The information presented here can be personalized to the user request depending on the extent of exploration.

Figure 5: A typical way to end the conversation with the bot after gathering information that is required for planning to visit a point of interest on the go. The amount of information required varies from user to user, but our bot should provide enough dynamic feedback to account for most experience levels.