Adding an AI date location feature for a better curated dating experience
UX/UI Designer
80 hours
Figma, Figjam
Capstone Project
OVERVIEW
Bumble is a woman-owned dating application that implements the feature of “swiping left/right” to match or reject potential match candidates. Though it gives users the opportunity to find new people in the area to date, it could be time consuming to research date locations or gauge each others’ interests.
the challenge
People looking for dates need a more convenient way to set up dates with one another while also ensuring safety.
the solution
Incorporate AI features to generate possible dating locations for dating app users.
Design a way for users to determine the safeness of a dating location.
problem
Dating others and getting conversations flowing is a difficult task for many. Setting up dates is another type of heavy cognitive load that can weigh down someone’s psyche, especially when it isn’t clear who should reach out first to set up an in-person date.
When is a good time to reach out, exactly? What is a socially acceptable time?
Being subtle and nudging might never be read right. Why not cut the middle man out now?
01 research
research 1 - user interviews
INFORMATION ON A LOCATION IS CRUCIAL
65% of users tend to look at reviews, photos, and parking prior. Little information of a location may be assumed to be untrustworthy.
LITTLE CONFIDENCE OF SETTING UP DATES
43% of users report that the biggest pain point for setting up dates was the constant back and forth conversations to decide on a first date location.
50% of users prefer to get their match's preferences when setting up a date.
key takeaway
Users prefer matching their location interests with their date’s interests. This way, users are able to easily find date places in as fewer steps as possible. Not only do the locations need to be similar, but also needs to be considerate towards travel time and safety.
research 2 - affinity map
I gathered findings and quotes from the user interviews and compiled them into an empathy map. From this, I analyzed and dug deeper to find out the general understanding of what the user desired. My findings were:
The most popular destination for dates are food based or activity based.
Safety is a big factor in choosing a date.
research 3 - competitive analysis
Even if Tinder has a “suggestion” feature, this feature is more so a curated stack of users who are interested in a particular suggestion. Users have to match with others by swiping left or right to reject or connect, however, users are locked only by people who are interested in this sole suggestion. Besides this, the other platforms appear to lack ways for users to set up interesting date locations in an area.
02 synthesis
SYNTHESIS 1 - PERSONAS
After gathering intel about the data received from the research phase, two personas emerged from the insights gained. The first persona is a tattoo artist who is crafted to serve as a guide for the goals and frustrations of established tattoo artists in the industry. The second persona is a customer, who was developed to foster a sense of empathy to the customer user base and to understand their objectives.
Emily and Sean both have similar challenges that can grouped into common themes: opportunity for choices, detailed information on locations, and lack of convenience. Keeping these findings in mind, I took into how the challenges can affect the design decisions.
03 IDEATION
ideation 1: user flow
To stand out of the competition, the use of AI will take into account the user’s preferences as well as their match’s preferences on where to go. This will address a few problems users already have when planning:
Indecisiveness: No more going back and forth while saying “I’m down for whatever you want to do”.
Ability to reach out without waiting for a socially acceptable time avoid having to hint who should ask first!
For the AI element of the project, I had to ensure that it was as easy to use as possible — given that using AI to generate items is a relatively new concept.
To understand the users’ preferences, I opted for a survey that users can complete during the sign up process. This is information would then be inputted into the AI system to better filter date places for the user and their matches.
ideation 2: digital wireframes
After deciding which direction to go with for each screen, I transferred the design to low fidelity wireframes.
Using the feedback I received, I made changes to the designs. I especially took into account the following criticism:
Asking the user if they want a serious type or casual type date may be a feature not used often. Users usually use the app to set up first dates then switch onto other platforms to continue communication e.g. texting, Facetiming, etc.
Introducing a new AI icon to an already crowded space may confuse users. Instead, put several of the functions under a button that opens a bottom sheet.
04 TESTING
usability testing
The wireframes gave way to creating high fidelities that took inspiration from the elements already provided by Bumble’s main interface. I made iterations along the way, such as changing the AI element box from being in the same text box as the private messaging text box into a bottom sheet. The rest of the task flow also happens on the bottom sheet in order to separate both the texting element and the AI element and to prevent confusion.
testing parameters
I built an interactive prototype for users to test remotely through the use of Maze, and unmoderated testing application. I kept close to the colors, vector art, and over all design of Bumble’s original UI kit. There were a total of 7 participants who have gone through three task flows:
Setting up a date location with a match
Leaving a review
Completing a survey during sign up
results
Feedback was gathered from the answers users provided from Maze and put into an affinity map where I analyzed the results. Though there was a 100% completion rate, I identified some critiques that I implemented into my reiterations. Click here for the full report.
Sharing the AI typing box with the same messaging typing box for the matched user was not preferred.
A few users reported they felt panic when they tapped the paper airplane icon that represented “send” after typing in AI recommendations on the messaging box itself.
priority revisions
I analyzed the pain points of the participants and made necessary changes. I selected iterations that would improve the over all experience and have the biggest impact while experiencing the feature.
Users were confused on where to submit their AI messages.
Reworked the screen to input AI messages in a bottom sheet instead.
When pressing “show me”, users are led to a separate screen.
INTRODUCING BUMBLE AI
Users are able to utilize the AI feature to find new dating locations that cater to the user's and match's preferences.
In order to report on the safety of locations listed, users can leave reviews for the locations they've visited. In the same vein, users can view reviews written by others for a peace of mind.
AI implementation is such a new concept in the design world that finding ways to streamline it as well as how to feed information into the AI system was a difficult task. Other similar apps do not offer this type of service and finding ways to implement it was a challenge. Because of this, I looked towards indirect competition to grab familiar elements for this process. Even if I had to accommodate these constraints, the over all design and palette Bumble offered kept my creativity going with no problems.
For this project, I focused primarily on the AI recommendation tool and the elements that surrounded it. Due to time constraints, the project was concluded once I made the necessary iterations. If I had more time to explore further, my next steps would have been:
Put in a user review page on the location detail feature.
Conduct another survey for further insight on users’ experiences with other dating apps.
Continue to test and iterate. I would have liked to study further on how users Bumble after going on that first date. Many users appear to migrate to text messages, but what about the users that choose to stay on the app?