Overview
The purpose of this project was to learn how to design and develop Voice User Interfaces. I used Alexa Custom Skill to design and develop the voice interaction.
Problem
Parents and college students, with their busy schedules, usually procrastinate to do their laundry as they find it time-consuming. Some usually forget to remove the clothes from the washer; others hate to spend time in front of the washer.
Solution
I present to you Smart Washer, an Alexa custom skill concept that enables Alexa to control your washer remotely, remind users when the detergent is expiring and order detergent from Amazon Fresh. Several products from Whirlpool and LG were introduced at CES, demonstrating the feasibility of the concept.
My Role
Full Stack Developer
Tools
Alexa developer console, AWS Lambda
Problem Discovery
User Needs
As per the Design Thinking methodology, I started by interacting with people, finding out their impediments. After empathizing with the people, I categorized their needs into three large categories: the ability to remotely control washer, get the real-time status of the washer, and controlling their washer without manual or physical interaction.
"I constantly forget to remove the clothes from the washer as I'm preoccupied with my children and work. I end up damaging the clothes."
– Ashwini the Forgetful
"I'm overloaded with work all day long. I do not have time to wait for the washer to complete."
– Amit the Workaholic
Why Voice?
The most natural and intuitive form of communication.
It is simpler to ask "Turn on my washer" while driving back home from work so that when the user reaches home the washer cycle is complete and the clothes are ready to be inserted in the dryer. In this way, the user will save his valuable time and this can be done without adapting to new technology.
Hands-Free experience
With the use of Voice UI, no physical interface is needed. This is especially helpful when your hands are occupied in work, for instance, driving a car or feeding your child.
Design
The main intention of this small exercise is to understand Voice Interface and Smart Home Technologies, and thinking boldly about future technology. As a bold and feasible move, the future smart devices can have the ability to order detergent directly through voice-over, provided recently by Amazon Alexa. Additionally, by ultra-sonic depth sensors, the quantity of the detergent can be measured which can provide a reminder when the quantity reaches a certain limit.
Dialog Flow
Taking the reference of the Alexa Skill Builder’s Guide, I was able to learn and implement Alexa Skill that will successfully engage a user in a seamless interaction by guiding the user with each response. Below are the two dialog flows that provide differences in the nature of tone and intelligent responses to concise utterance.
Design Considerations
1. Recommendations: Provided user with a user-friendly recommendation so that the user will remain engaged and reducing the effort to think on what next to ask.
2. Follow-up prompts: Usage of questions to provide clear indications that a response is expected, without explicitly telling the user.
3. Natural tone: Responses provided in such a way that it feels like we are having a conversation with a human instead of a machine.
4. Avoidance to information overloading: Precise information is provided to the user without any unnecessary data. Simplicity is taken into consideration where ever possible.
Challenges
Synonyms
Same ideas can be expressed in multiple ways, without speaking the exact same words. Hence, multiple synonyms were created for each utterance and feed to the prototype.
User wants to know the status of the appliance
How's my washer doing? What's my washer doing?
How much time left on my washer? Time left.
My washer status? Washer status update.
User sets a cycle on the appliance
Set my washer to the usual cycle. Set washer to Quick wash.
Set washer to default cycle. Start my washer the usual way.
Set my favorite cycle on the washer. Start the routine way.
User sets a reminder
Remind me before the spin cycle. Notify me to remove the clothes from washer.
Remind me after spin complete. Ping me once washer completes.
Share an update after each cycle Inform me when detergent is low.
Speed Breakers
Sometimes users may not follow the happy path to answer the question. They may come up with an unexpected question or option. Hence, I listed out a couple of such unexpected responses by the user and enabled the skill to handle such situations to maintain the conversation.
Other Possible Responses
I didn't get that. No need to provide status of the washer?
Can you repeat that? Provide time left in seconds instead of minutes?
Remember my choices. What's wrong with my washer?
Lessons Learned
The synonyms and the speed breakers or unexpected responses were not exhaustive, nor was I able to incorporate every situation in my design. However, it was a good thought to exercise, that pushed me to think more about how to design for voice, with all the unexpected inputs from the user.
Refining Interactions
I then thought about different use cases based on the unexpected responses and our initial user needs and refined my scenario by adding the following two paths:
Reminder For Servicing Ordering Of Detergent
Note that, to enable the functionality to order detergent, users will be required to link the skill to their Amazon Fresh account, and give the skill access to their mailing address. They might also need to manually select the detergent they want to buy the first time they use this function, so that, Smart Washer skill can remember their preference and order the same detergent the next time it is needed.
Future Steps
There are a lot of things that need to be improved by more iterations and some of them can be:
Initial Experience
A lot of the interactions were build based on the assumption that the user has already gone through the manual for linking the smart washer with their Amazon Alexa application and the skill. The initial setup experience can also be provided with voice that still needs further work.
Personalized Experience
With time, Smart Washer skill should learn a lot about the user's routine step of commands and the time at which the utterances were given, and provide experience tailored to these preferences (e.g. when to give the reminder of the washer status to the user and what servicing agency to call). I still need to work out these details to make the experience more user friendly.
User Testing
Even though there were some users to refine my design, more comprehensive user testing sessions are required to validate whether the design helps better solve our user's problems or not and whether the interaction makes sense for a variety of users in different contexts.
