GOFIND mobile app - Case study
Designing for Artificial Intelligence
UX/UI Designer & Visual Designer
(Partly alongside 1 UX/UI designer)
Involved in all parts of the design.
Collaborated with Data Scientists,
Deep Learning Engineers and Developers.
The CEO was the main stakeholder.
Shipped iOS and Android app
Mobile app designs
Customer Rewards program
Mobile push notifications - design & content strategies
Adobe Creative Suite
Pen & Paper
GoFind is an early stage fashion startup.
When I came to GoFind the CEO wanted to ship a MVP to the app stores.
Giving millennial shoppers the ultimate shopping experience, on their phone.
Helping users, where ever they are in the world, find products for sale by any brand, in their price range.
The scope of the MVP was constantly changing and growing.
Having to design without existing users, since the product wasn’t launched.
When I came to GoFind they have had designers working on the product prior to me. I was tasked with taking the existing designs to the next level in order to be able to ship a minimal viable product (MVP) to the App stores as soon as possible.
Reviewing the inventory
I started out with reviewing and organizing the design assets to see what we could use moving forward and what needed to be created from scratch.
One of the first things I created was a style guide since that didn’t exist. It would be helpful to keep a consistent design and ease the work for the developers if they had the assets and measurements all in one place. As more designs were being created we could make sure to follow what was set for the brand in terms of colors, typeface etc.
PERSONALIZED FEED IMPROVES WITH USE
The product feed is where the users spends most of their time, discovering new products, "like" items, shop and discover brands. My idea was to create a clean design were the products would have the center stage.
The logic behind the feed - How to personalize the experience?
In order to customize the experience I held a brainstorming session with the other UX designer on the team. We came up with a formula for how the A.I. would work.
The personalization starts out with the onboarding where the user answers a quiz. Depending on the answers they gave they were given a persona feed. The algorithm would later be refined by using the GoFind app, we had given different weight to certain actions like searching certain products, liking a product, clicks and so on.
- Crisp white background to let the items speak for themselves and minimize visual clutter.
- Prices - My research showed that millennial shoppers are very price sensitive. Therefore it was important to show the prices in the feed.
- Staggered layout to create visual interest and a feeling of infinite scroll.
PRODUCT PAGE Feed
When a user click on a product on the feed it will take them to a product page where they can take a closer look at the item and get a short information about the product. The click also initiates a new search for similar items (using the GoFind visual search technique), to what they just clicked on. With this feature the user gets more options in their search.
- Large product picture to showcase the item.
- The clear and large call to action button and a link to the original posting/original shop.
- Similar items below the featured product for more options.
VISUAL SEARCH Engine
The GoFind app has a visual search feature which is powered by artificial intelligence. Since GoFind is an aggregator the A.I searches an enormous catalogue of items for sale by almost any brand in the United States.
- The user will snap a picture of an item they would like to find.
- Gofind searches for similar items for sale.
- The results are presented in a product feed.
Working for GoFind gave me an opportunity to work on many different projects, with exposure to a variety of problems and techniques. I learned a lot from working closely with A.I. engineers, data scientists and developers.
Being at a startup is sometimes like throwing yourself into a whirlwind. I had lots of fun and I landed on my feet as a more experienced designer.