From deep research and interviews we discerned three main problems; Security & Privacy, Voice Assistant, & Speaker Quality.
Security & Privacy is at the forefront of everyone’s mind, especially when using a smart speaker; With the Voice Assistant there’s another problem, sometimes it does not quite understand what you’re saying…even after you’ve repeated it 4 times. There’s also the issue of it being too robotic and artificial, and especially now in the time of Covid it’s important for people stuck at home potentially alone to have an outlet to talk to and hold conversations with.
Going forward, we asked ourselves two questions: How do we implement security and privacy transparency for people to feel secure and safe connecting their device to the IoT?
How do we make our assistant more responsive to low utility requests, more conversational and less robotic while also staying ethical?
HIGH LEVEL TIMELINE
2 weeks to research and design
MAKE OF THE TEAM
Pia Kahl, Ryan Rosales, Jaskaran Singh Sondi
Solving for the problem of privacy and security in IoT and Smart speaker
I was primarily the lead researcher, interviewing and surveying users of current IoT and Smart speakers to identify problems.
I researched IoT (Internet of Things) in depth as well as performed a competitive analysis of brands, dove into issues with Smart tech and Smart homes. During this two week research focus, I conducted ethnographic interviews, surveys and created personas. Through my research I identified the 3 core problems as well as solutions before we began iterating our app.
For the creative process I created the name, Synneko as well as the Voice Assistant, Iris. I assisted in building out the design system, created the pitch deck or our final presentation and was the sole presenter.
These Smart speaker designs were created by team member Jaskaran Singh Sondi for our final pitch and prototype.
UNDERSTANDING THE USER
After analyzing quantitative data and conducting qualitative interviews we uncovered that the length and usability of the mobile collections page was driving average order value and conversion rate down, and with over 86% of visitors coming from mobile this was something we focused on first.
During the research phase I interviewed current users of Smart speaker and IoT, a high level scope focused on what IoT or products they currently own, how they use it, why they use it that way, and what they do and do not like. We identified that our target audience is between 25-50 years old, but there are some Smart speaker users that are older than 50 who mainly use them as a timer and for getting quick information instead of as a music speaker or as an IoT.
By creating our own proprietary cloud service, Synnekt, people that currently use other Cloud based systems can feel more secure in their choice and finally feel comfortable enough to connect their device to other Smart products in their home allowing the speaker to be utilized to its full capabilities rather than just as a timer.
Synneko Cloud (Proprietary Cloud)
Data exportation by the user using 2 level authentication to increase data transparency
As designers we need to consider every variation of intent, utterance and slot of a voice command.
To create a more responsive and conversational Voice Assistant we would utilize machine learning algorithms to ensure that the assistant can keep up and adapt to situations, continue learning and understanding voice requests and conversations.
Building a Dialogue Flow
Consider every variation of intent, utterance & slot of a voice command
Utilize machine learning algorithms to ensure that the assistant can keep up & adapt to situations
Systems thinking allowed us to view the broader viewpoint and interactions of people, stakeholders, designers and researchers.
Design thinking can often times have too limited of a scope, and for this project we wanted to think about the business and user goals and needs.
Iris is the Intelligent Rad Interactive System.
We wanted Iris to be more adaptable and flexible in conversation, this was especially important for us to think about as during Covid-19 millions of people are at home, and some without anyone to talk to.
A more responsive voice assistant
We created a dialogue flow to better understand how people could interact with Iris and to make sure that we create a Voice Assistant that can understand the variances in our speech.