2 weeks ago, we prepared for the meetings with the mentors.
The preparations included a discussion of Dapi's design principles and raising questions that could guide us toward a more correct and appropriate solution.
Among other things, we have also developed a pitch to present Dapi.
Here is our pitch:
Our group’s aim is to improve the experience of the driver and passengers on city bus drives, by improving the interactions between the driver and his passengers.
The problem we chose to approach arose from interviews with many drivers - drivers feel transparent, a feeling that is typical among service providers. Politeness and courtesy towards service providers in general, and bus drivers in particular, has declined over the years. We wanted to improve the interaction between the passenger and the driver, with the understanding that improving the experience of boarding the bus will affect the whole trip.
When a passenger comes to the driver to pay for the ride, Dapi will be able to identify words that indicate politeness such as "Good morning, hello, thank you" etc. Dapi will engage with the passenger, with a cute motion or showing something on its screen. This will encourage other passengers to interact with the driver.
During the meetings with the experts, some points were repeated. Some of the mentors were concerned that the interaction will be between the robot and the passenger, eliminating the driver from the interaction - exactly the opposite of what we want to do. Another problem that arose was how to measure kindness and politeness. Perhaps it is too mechanical count words like ‘hello’ and ‘thanks’ as courteous.
We were encouraged to look into gamification in order to get people excited about it. This could mean adding an app that can give stars to passengers who respect the driver, or relaying messages to the driver showing the feedback that we received from passengers on the bus. Dapi could perhaps indicate to the driver through colors what the general feeling is on the bus. Green if the passengers are happy, red if not.
We were also advised to humanize the driver to the bus passengers. If the passengers learn more about the driver, for example his hobbies (Get Taxi style), they will be better at empathising with him.
The most significant point that was brought up with the mentors was timing. We were advised to spend more time thinking about, and reevaluating, whether the timing of our interaction makes sense. Our plan was for Dapi to interact after/during the interaction with the driver. We were encouraged to think more about whether this should happen before, during or after the interaction.
Last week, after meeting with the mentors, we discussed all of this between ourselves. We decided that we do not want the interaction to start before the bus arrives; we want passengers to see the bus as the bus driver’s arena, so starting beforehand would not be helpful. We want the interaction with Dapi to start as soon as the bus pulls up, as the passengers begin boarding.
Dapi will encourage the interaction before it happens, by moving around. When a positive interaction occurs, Dapi will have a happy response (a change in color perhaps - nothing too distracting from the conversation). Dapi will count the amount of interactions, and will give positive feedback to the driver after a certain amount. Dapi will recalibrate itself depending on the driver - if for one driver 10 interactions happens quickly, Dapi will only give the positive feedback after 20. If the driver hardly gets 10, Dapi will bring it down to 5.
In this way we will encourage the passengers before they have approached the driver, as well as giving the driver a gamified experience.