An Apple Mail redesign that features more deliberate user participation in machine learning through microinteractions.
OVERVIEW We designed an update to Apple Mail’s sorting abilities by making AI decision-making more apparent and reactive Instead of proactive.
PROBLEM
OVERVIEW
How might we empower people to keep their inboxes organized more easily?
INSIGHTS Existing email clients have ways to build “Smart” mailboxes which filter emails based on certain criteria but based on testing, we found that people were frustrated with trying to figure out and anticipate what might be useful for them in a mailbox. We also found that automatic sorting, like Gmail’s tabs, can make mistakes and it wasn’t clear how to fix.
SOLUTION A responsive mailbox that makes suggestions to sorting and invites the user to correct mistakes as they happen.
An updated Apple Mail which puts sorting at the forefront and empowers AI to be more proactive and visible to the user.
BRIEF
PROCESS
The most straightforward occurrence can make or break your day - dropping a hat in the mud will set a chain of unfortunate events. A smile from a stranger will brighten your whole day. The interactive mundane has this power over us.
In this project, we redesigned users' interactions with desktop email clients. We focused on the power of microinteractions to make those digital chores easier, simpler, more enjoyable - and less noticeable. The goal was to move the task to the periphery and to bring forward the delight of being able to communicate instantly across the globe.
CONTEXT
PROCESS
AI has become ubiquitous. We see it in our inboxes-- pre-sorted folders put our emails in Spam, Updates, or Promotions. But, as consumers, we don’t really know how AI is making decisions. So when something goes wrong, people are left disempowered with no way to fix it.
As a team, we compared email habits. One of us had 19,573 unread emails in their inbox. This is a failure of the tool. So, what isn’t working?
RESEARCH
PROCESS
We created a SWOT analysis to better understand precedents and competitors. This was our primary form of research for this particular project.
SWOT ANALYSIS OF EXISTING EMAIL SERVICES PROPOSAL
PROCESS
Create more deliberate user participation in AI training
METHODS
PROCESS
Fine tune pre-sorted folders through Machine Learning using natural language processing (NLP) models
More pre-sorted folders to choose from
Periodically and automatically delete emails and parse through emails for expiration dates
PROTOTYPES
CLICK IMAGES FOR FIGMA PROTOTYPES
Setting up FoldersCourse CorrectionDelete SettingsREFLECTION
During this project, I loosely adopted the Extreme User research method which chooses a user with extreme needs, which can often be a stereotype of a person. By focusing on a user who doesn’t have the time to understand features or to sift through and organize their inbox, we came up with a way to have the sorting process front and center and in small bites to avoid overwhelm.