MANDY



Team



Roshini Deva, Lilit Sargsyan, Sugandh S. Bagaria, Chia-Yun Tsai, Aditya Deshmukh, Suhaib Mutawakel, Muhammad Chambers



Role



UX Researcher / Designer



Academic Advisor : Feng Zhou, PhD, Assistant Professor at University of Michigan-Dearborn
Period : September 2021 - December 2021



Overview



Mandy is a design concept for a voice assistant that relieves drivers from having to interact with their cell phones by reducing manual and visual interaction with them while minimizing their cognitive load.

In the context of the Foundations of Human-Centered Design and Engineering graduate course at the College of Engineering and Computer Science at the University of Michigan-Dearborn, my team developed this student team project.



Context



Official statistics indicate that using a cell phone while driving is a leading cause of distracted driving. Thanks to recent advances in technology, we can now engage with social media, music apps, and various messaging apps on our smartphones, resulting in an increase in accidents on our roads.

The idea of voice recognition as a solution to the distracted driving made sense to us when we started our research on voice assistants. After reviewing the relevant literature, analyzing customer feedback on the design interventions currently in place to minimize distraction and comparing the popular voice assistants as they relate to their texting features and integration with the most frequently used messaging apps, we found an outstanding problem that needs to be addressed: cognitive distraction. Despite the fact that current voice-based apps reduce visual and manual interaction with the device, they tend to increase cognitive load on drivers due to poor conversation design, difficulty understanding foreign accents, as well as poor integration with third-party features.

As a solution to the distracted driving problem, we proposed a design that would address our findings.



The Design challenge



We defined our design challenge in light of the initial findings as follows :

“How might we reduce the cognitive distraction drivers are exposed by cell phone usage?”

Two possibilities were considered as a possible approach to solving this design challenge :

1) To design an interactive voice user interface optimized for the context of driving; the interface would be seamlessly integrated with the functions most often used while driving, would involve the least amount of queries to accomplish the action, and would require minimal visual interaction; and

2) Design an adaptive user interface that will restrict cell phone usage under high traffic conditions.



The Process



Empathize



Study the behavioral patterns and pain points of drivers interacting with their cell phones while driving



Define



Outline the needs and goals we want to meet with our design



ideate



Sketch possible solutions that can address the problem of distracted driving based on the needs identified



prototype



Develop a voice user interface design approach that reduces unnecessary cell phone interaction while maintaining functionality



test



Test the usability of created design



User Research



I conducted a series of interviews to gain a thorough understanding of the problem and to identify the target users. The interviewees were chosen through a screener survey, which included, in addition to standard demographic questions, questions asking participants to briefly describe the scenarios in which they use their phones while driving, as well as the difficulties they encountered when using voice assistants. Eight participants were chosen for the interviews from among the 20 responses received.

Our next step was to create a questionnaire to ensure consistency across the eight interviews. The interviewees were asked to recall situations in which they had to read and send texts on various messaging apps, receive and make calls while driving, or any other scenario of cell phone usage while driving, and to describe the difficulties and coping mechanisms they encountered, as well as the accompanying emotions. In addition to providing insights into anticipated scenarios, the interviews assisted in revealing a new scenario in which users must use their smartphone while driving: making restaurant reservations and ordering food delivery. This discovery would guide future work on developing a new feature for the mobile application.



Research Findings







In order to organize the information collected from the interviewed participants, the affinity diagram method was used. The research findings were then consolidated using the miro software. A different color was assigned to each of the eight interview participants on the miro board. The sticky noted were then organized by the themes identified and each group was labelled.

The following major themes were identified from the interview data pinned to the miro board :

1) Connection

2) Full access to the phone's functionality and

3) Safety



Mainstream user Persona



Robert Lin



“I would like to stay connected and have full access to my phone while driving but I need to feel safe too. It would be great to have good voice control that would minimize unnecessary interaction with the phone while keeping the functionality.”


Age - 31 years old

Languages - English, Chinese

Education - MS, Software Engineering

Occupation - Works at a tech company

Location - Bay Area, California

Marital status - Married, 1 child


Behaviors

* Drives daily

* Connects his phone to the car and checks quite often

* Rarely uses a voice assistant

* When driving, reads text messages and takes calls from wife and other family members

* Sometimes handles urgent business correspondence in his car

* Accepts calls by pressing a button on the steering wheel Rarely responds to text messages when driving

* Loves to listen to music

* Regularly uses navigation


Pain points

* Does not feel safe when using the phone while driving

* When there is no passenger, reads text messages on his own; waits for a Red light or light traffic condition under 45 MPH to read

* During emergency, types short standard answers (1 or 2 words)

* The voice assistant does not understand his accent

* Repeats the commands several times, looks at the screen and touches buttons

* Gets distracted by unimportant messages and calls

* Does not want to miss calls from his family by putting his device on silent mode.


Goals

* Navigation

* Listening to Music

* Communication with family

* Handle urgent business communication


Needs

* Safety

* Connection

* Language Functionality


Likes and Dislikes

* Prefers the car display due to its size and location

* Big notifications popping up on the screen and the sounds are distracting


Expectations from the product

* Good voice control for communication, navigation and music

* Join zoom meetings, facetime and respond to emails

* Make restaurant reservations or order food through voice control from the car

* Wish to use the phone only for communication with selected contacts while driving



Extreme user Persona



Alice Cole



“Safety should be prioritized above anything else. Usage of various messaging apps while driving is distracting. It should be restricted. When I am driving, I need my phone only for receiving calls from my family. It would be great to have a simple voice control that would allow to handle calls hands free.”


Age - 48 years old

Languages - English

Education - BA, Accounting

Occupation - Works at a bank

Location - Dearborn, Michigan

Marital status - Divorced, 3 children


Behaviors

* Drives daily

* Does not connect her phone to the car

* Has a voice assistant installed in her phone but does not use it

* When driving, ignores incoming messages

* Never texts while driving

* When driving, answers only calls because she feels they are urgent

* Receives calls mainly from her parents and children, when driving

* Accepts calls on her phone

* Regularly uses navigation

* Uses her phone for personal rather than business communication


Pain points

* Does not feel safe when using the phone while driving

* Used to text while in the car until she witnessed an accident

* She is not tech savvy; has to learn how to interact with a voice assistant

* Find interaction with the voice assistant complicated; besides, she has to speak a lot herself Finds phone usage while driving to be distracting and unsafe

* Needs her hands on the wheel while driving but her car is old and she cannot accept calls by pressing buttons on the wheel


Goals

* Navigation

* Communication with family


Needs

* Safety

* Availability (for selected contacts)


Likes and Dislikes

* Prefers to communicate via calls

* Messaging apps are distracting


Expectations from the product

* Simple voice control to handle calls hands free

* Restricting the usage of messaging apps
* Wish to use the phone only for accepting calls from selected contacts while driving



Story Board



Based on the user research, a brainstorming session was organized for creative solutions. As a more tangible vision of the design emerged, we drafted seven futuristic scenarios showing a persona interacting with the application. Illustrations of the drafted scenarios were created with storyboards.







Sketches



A lack of resources and limited time led us to focus on developing three main features

* Setting up the application

* Making restaurant reservations

* Handling business emails


The next step was to create paper sketches and low-fidelity digital screens for the selected functions in Adobe XD. We then presented the sketches to users to collect feedback.


Sketches enabled the team to capture every idea that came up and, based on each others ideas, sort out the features most suitable for prototyping at an early stage.






Low-Fidelity Prototype



An additional brainstorming session was held after gathering feedback. Applying the design funnel technique to the feedback received, the ideas most relevant for our design challenge were picked, and low-fidelity prototype was created in Adobe XD.


The low-fidelity prototype illustrates the steps for setting up the account for the user.






Choosing the Name



When we were thinking of a name for our driving assistant, our focus was on our target audience and the right language to reach them. We wanted drivers to see it more as a companion than just a mobile application. Our app needs an image of a caring, empathic, trust-worthy young woman in her late 20's or early 30's, the same age of out target drivers. One of out teammates came up with the name "MANDY" which seemed suitable in this sense. It was also appropriate from a linguist point of view, as the app could be configured to pick up the frequency of the name being used as a trigger word.





Color palette & Font



The color palette of the interface was chosen in accordance with the design guidelines for android auto. Parallel to the dark mode version, a light mode version was designed as well. However, we chose to stick with dark mode to minimize distraction.






High-Fidelity Prototype



We converted the low-fidelity prototype into a high-fidelity design by using Adobe XD software, which is capable of creating voice-over prototypes.


Below given is an initial version of the high-fidelity prototype showing the pages for setting up the app, viewing the home page, and driving while using the app.



Usability Testing



Five users selected from the target audience were asked to test the high-fidelity prototype. We developed a usability testing questionnaire and scenarios for testing the tasks such as setting up the application, replying to an email and reserving a table at restaurant.

In an ideal testing scenario, users would have interacted with the app while driving, but such contextual inquiry testing was not possible because of technical restrictions. Thus, usability testing interviews was conducted both online and in-person. The obtained feedback generated valuable insights for iteration, and we made a number of design improvements based on the results.





Iteration



1) Setting up the account was simplified :

  • We removed the while driving page and redesigned the application to run in the background of the phone until activated
  • A number of screens were removed from the voice assistant, the reasoning being that it should act upon any command it receives on the spot, including creating personalized commands, selecting commands to display and setting up routines
  • In the new setup process, the user can land on the home page screen after setting the voice model : on home page, user can set up the high distraction level rules, select the apps to be used while driving and add the favorite contacts to allow while driving


2) Making reservations at restaurants was made easier by :

  • To accomplish task with as few queries as possible
  • The conversation was redesigned for different scenarios of making a reservation, including : location and suggesting a list of restaurants, placing a call to the restaurant for the user to make the reservation, calling the restaurant and booking a table automatically

3) In the additional settings, the user can now choose whether or not the conversation transcript is displayed on the screen







Interactive prototype



The video showcases the final conversational prototype demonstrated by one of my teammates



Key takeaways



As a student, I completed my first UX project, Mandy. As a result, it provided me with hands-on experience in carrying out a thorough human-centered design process. The following are some key takeaways from this project:

  • Sketching and testing a variety of ideas early in the design process provides richer insights and opportunities for final design, saving time and resources later in the process
  • Extreme user cases can reveal the latent needs of mainstream users: designs that typically serve a small group of users with special needs may turn out to be the best solution for the majority of users in specific conditions
  • Contextual inquiry type usability tests and interviews are one of the most reliable and effective methods of understanding the design problem, defining user needs, and evaluating the designed solution because what users say they do or would like to have does not always correspond to what they actually do or need. Observing users interacting with the product in a natural setting gives a deeper understanding of the real problem and helps validate the designed solution


Future possibilities



  • Consider recruiting a larger pool of participants for user research
  • Conduct field studies to observe participants actually interacting with their smartphones while performing the primary task of driving
  • Spend more time on ideation by sketching more ideas and testing them with a larger number of users
  • Conduct multiple rounds of usability testing and iteration
  • Complete a summative evaluation for collecting quantitative data in addition to qualitative data to measure the overall success of the final design in terms of addressing the design challenge.


Get In Touch