We live in large city and commuting by public transit is a major (and often not the most pleasant) part of many people's lives. What's more frustrating than waiting for your bus to take you to the subway station only to then find out that there is a major service interruption on your subway line and you're going to be late for work? What if we created an Alexa Skill which could provide you with all the essential information about your commute while you are busy doing other things in the morning, like packing your lunch?
What it does
My Transit NYC provides you with a quick personalized update on the status of your daily commute. This includes the next arrival time of your transport as well as updates about any potential delays with information about service alerts sent directly to your phone. All you have to say is "Alexa, open My Transit".
How we built it
At first we did a couple of rounds of user interviews and testing with low-fidelity prototypes to understand what needs commuters have and then map them to Alexa functionality and pick the use case. Then we did role-plays and more user testing to define the conversational flows: the happy path for daily use and onboarding, and then added the "conversation repair" flows. In the meantime we also did engineering research to ensure feasibility of the desired ux. After we were done with these preliminary steps, we proceeded to the "build" phase.
Challenges we ran into
From the UX and product standpoint the biggest challenge was the onboarding flow where we had to get the start and end point addresses from the user and also to learn their preferred commute route. Providing this information by voice is a lengthy process, so we split it up into steps and also included some expectations setting in the beginning, so that the user was aware that onboarding will take some time, but they would only have to do it once.
From the engineering standpoint the key challenges were:
- Transit data is a world by itself as each and every transit company uses its own format and designation for the different transit (for example some cities will call the outdoor transit tramway and we call it street car in Toronto).
- Another challenge is in capturing the data address, it is pretty hard to manage conflict in the same city, we solve it by having the boroughs as a conflict solver and by leveraging the
dialog delegateoffers by Alexa in New York but moving forward we will work on a deeper conversation repair to address this.
- Lastly maintaining the session over a back and forth conversation is really tough and makes conversation repair almost a nightmare.
Accomplishments that we're proud of
We think that we were able to identify a good use case and also design the onboarding flow in the best possible way and the skill has potential to scale to multiple cities.
What we learned
We developed and refined a set of techniques and tools that we use for voice-first products and which allow us to uncover user needs, find the best use case for voice and design the conversation flow.
What's next for My Transit NYC
Currently the skill supports only NYC area and Toronto, we are planning on expanding it to a wider number of cities as well as including other personalized updates commuters will appreciate like weather updates for the times they will be commuting to and from work. We will optimize the skill for Echo Show and Spot and then we are also planning to create a similar skill for cyclists specifically.