Taste.ai
Taste.AI offers a user-friendly experience developed following the UX process. The goal was to design an AI-driven food or recipe recommendation app that makes deciding meals effortless. The app provides quick, personalized suggestions by considering dietary preferences, available ingredients, and time limitations.
Role
Solo project
Duration
12 weeks
The Problem
In 2022, there were approximately 37.89 million single-person-led households in the United States. The number of single-person households is increasing, and when living alone with a variety of ingredients, it's easy to generate a lot of food waste. Naturally, people tend to buy less.
Users face difficulty in selecting meals due to a combination of limited ingredient options, specific food preferences, and a lack of appetite. They need a fast and personalized solution that suggests meals based on their preferences and what they have on hand.
FEATURE 01
AI Chat
Provides AI-generated food recommendation personalized suggestions, ingredient-driven recipe ideas, and restaurant recommendations based on location and preferences.
FEATURE 02
Ingredient Scanning
Users use the camera to scan their surroundings or the contents of their refrigerator.
FEATURE 03
Customization Based on Preference
Analyzing user preferences, dietary needs, and ingredient availability, offering tailored recipe and restaurant suggestions.
Design Process
01/
Empathize
· Competitive Analysis
02/
Define
· User Journey Map
· Persona
· Affinity Diagram
· User Quotes
03/
Ideate
· Idea Brainstorming
· Sketch
· Information
Architecture
· User Task Flow
User Research
17 Participants
· User survey : 17
· Interview : 6
Requirement
A people who live alone
Demography
· 10s~50s ages
· 50% Male and 50% Female
User Quotes from User Interview
"I usually check my pantry to see what I have and consider my dining options based on its contents. I like to choose my meal based on what's available in my pantry."
Brian / Lawyer
Too hungry to think deciding what to eat can be challenging
sometimes it takes over an hour due to my indecisiveness.
David / IT consultant manager
“As a vegetarian, the food you can eat is limited. I want to find more options.”
Darjar / Marketing director
“I’m on a keto diet. recipes and ingredients are limited for keto so I want to find some new recipes”
Sally / Art director
Affinity Map
Key Finding
of participants eat out less than three times a week.
71%
of participants took 15~20 minutes to decide what to eat, 15% took 20 mins to an hour.
75%
of participants discover new recipes to try through the social media.
70%
of participants expressed it’s helpful to look at menus or food options online.
90%
of participants expressed they don't make a weekly meal plan.
Persona
User Journey Map
Problem Statement
01/
How might we help users spend less time thinking about what to eat?
02/
How might we help users in removing unwanted ingredients from their preferences or recipes?
03/
How might we help users find recipes with their available ingredients?
Solution
13/
AI Chat
This flow allows users to ask anything and engage in a conversation with the AI.
02/
AR Scan
This flow allows users to ask anything and engage in a conversation with the AI.
03/
Customization
This flow allows users to ask anything and engage in a conversation with the AI.
Information Architecture
User Task Flow
Three main user neeeds
User Testing and Major Improvement
Ai Chat
Could I ask a little more detailed question?
Added a detailed question option to help users understand its use for the first time. Also, moved it downwards and provided more options.
Search Page (Filter)
I’d like to see the search bar first.
Each category has so many ingredients that it's hard to see at once. The scrolling is also too long.
Moved the search bar to the top and now displays popular ingredient tags first. If the user presses the '+ Add Ingredient' button, they can see all the ingredients in the bottom sheet. Additionally, a recently added ingredient section has been added.
Controlling the time accurately is difficult. Is there an easier way?
Switched to buttons and provided three options for time: under 30 minutes, under 1 hour, and under 2 hours
I didn't notice there was an expansion option there.
Moved the button next to the last one, and there is a "+ More" button for expansion if needed.
Recipe Detail
How to recognize vegetarian food here?
Added some tags for food preference, cuisine, and difficulty level.
It's hard to read because there's so much content.
Divide content using tabs (ingredient, instructions, Nutrition facts) so users can quick access and added button to view full recipe.
Style Guide
Reflection
What’s Next?
· Develop screens for the shopping list page and profile page
· Proceed with the development of a desktop prototype
· Develop AI-based meal planning for families
· Customize meal plans for individuals with specific dietary requirements
· Implement delivery for groceries and food items
Lessons Learned
I knew that the design process was crucial for creating a good user experience, but this case study really drove that point home. Initially, I had some ideas about what the final design of TasteAI might look like, even before I began working on it. However, everything shifted once I conducted research and defined the project more clearly.
It was nice to be able to build up knowledge about AR and AI while doing research