CulinAIry
Cook with what you’ve got!
Merging technology with culinary creativity, CulinAIry is an innovative recipe generator app designed to simplify the cooking experience.
This project was part of a 15-30 day challenge to create a Minimum Viable Product (MVP) with TechLeap, and my team, which consisted of a Product Manager and two other UX Researchers.
As the sole UX Designer, I lead the team through design exercises, ranging from sketches to final designs. The experience was built on Glide to utilize their no-code/low-code tools.
View our presentation here!
Problem
Have you ever struggled to find recipes that fit your preferences, restrictions, and ingredients?
You're not alone—many face decision fatigue, wasted time, and lost confidence in the kitchen.
Solution
Develop a recipe generator app that uses user input regarding food restrictions and available ingredients, leveraging AI to create personalized recipes.
Research
Due to the unique structure of the program, secondary research was prioritized over first-hand research, and our team was tasked with utilizing AI tools such as Perplexity and ChatGPT to practice crafting prompts and conduct market research.
Through secondary research, we learned:
A majority of users are 25-44 year olds, making up 60% of users overall
Ideal users include busy professionals, budget-conscious consumers, novice cooks and family meal planners.
Meal recipes are not personalized with easy-to-use filters where the focus is often on health or recipe sophistication.
Personas
Busy Professionals
Need: Convenience and time-saving solutions for meal preparations.
Characteristics: Working long hours, limited for grocery shopping, looking for quick and easy recipe ideas.
Budget-Conscious Consumers
Need: Cost-effective meal solutions that utilize ingredients already available at home.
Characteristics: Managing tight budgets, looking for ways to minimize food waste, seeking affordable and versatile recipe ideas.
Novice Cooks
Need: Simple and beginner-friendly recipes with step-by-step instructions.
Characteristics: Limited cooking experience, seeking easy-to-follow recipes, looking for guidance on ingredient substitutions.
Family Meal Planners or
Single-Person Meal Plan
Need: Recipes suitable for feeding 1 or more people.
Characteristics: Managing meal planning for multiple family members looking for wholesome and satisfying meal options.
Brainstorming
Once we had gathered research from other apps, we came up with 3 How Might We’s to shape our design process.
How might we…
streamline the user onboarding process to reduce fatigue?
How might we…
track post-cooking experience and our indicators of success?
How might we…
build trust with users that their individual preferences are being addressed and their feedback is being taken into account?
Understanding User Expectations
Survey results revealed users’ three key pain points, and we were challenged to come up with three ways to improve the user experience.
Personal Challenges
Novice cooks commonly face challenges such as limited meal options, decision fatigue and inefficient routines, which hinder their cooking experience.
Customization
Onboarding asks the essential questions that customize the app experience to user’s unique background and experience with cooking.
Adaptability and Waste Management
Novice cooks struggle with adapting recipes to changing circumstances and managing food waste, which can lead to inefficiencies and environmental impact.
AI-Generated Recipes
Users will receive personalized recipes based on the information input during the onboarding process and the recipe generation process.
Emotional and Psychological Aspects
Emotional factors such as lack of confidence and heightened stress significantly impact novice cooks’ overall cooking experience affecting both enjoyment and proficiency in the kitchen.
Three Recipes Max
By providing users with only three recipes, we eliminate the decision fatigue that can accompany searching for recipes online.
Sketch to Screen
For this project, we sketched three key pages:
Onboarding
Homepage
Meal Selection
Our sketches were grounded by two sprint questions:
Can the app provide personalized recipes?
Will the onboarding be able to accurately reflect the true inputs to the app?
Exercises: Solution Sketch, Art Museum, Heat Map, Straw Poll, Speed Critique, Super Vote
My team met to share our many wonderful, creative ideas.
Due to the interest of time, we ultimately decided to narrow our focus to practical MVP concepts for this sprint:
Users can input their dietary needs / restrictions, meal type, level of cooking skill and ingredients available.
To avoid overwhelming our users, we will limit AI’s recipe generation to three responses.
To save time, we will focus on designing only for native mobile devices. Other devices to be explored and designed for during later design sprints.
Storyboarding
Storyboarding allowed our team to take a step back and prioritize which proposed features were most important for our MVP:
We determined that the first screen should include questions usually found in an onboarding flow such as dietary needs / restrictions, meal type, level of cooking skill and ingredients available.
While our other ideas were great, due to time constraints, we decided that we would highlight only two key features:
Users can input their dietary needs / restrictions, meal type, level of cooking skill and ingredients available.
With this information, three relevant recipes will populate.
Our initial storyboard flow was convoluted, but this exercise helped us better understand our users' perspectives. As a result, we revised the storyboard to create a simpler, more intuitive path.
User opens app and creates an account
Users goes through the onboarding flow and adds info
User chooses a meal type and enters ingredients to generate 3 recipes
User is able to follow along with instructions
Recipe is complete. User is rewarded with confirmation pop-up
Target & Recruit
We conducted interviews with three participants to help us fill in gaps in our research and refine our MVP, to ensure it effectively addresses the following challenges:
01. Will the user onboarding reflect the true inputs for recipe?
Answer:
It does not reflect the cuisine or the types of food they like to eat
Very few of the dietary restrictions / allergies were being filled out
User research needed to understand other inputs needed
02. Can the app provide personalized recipes?
Answer:
The app provided personalized recipes based on ingredients
It did not account the cooking level and provide detailed enough steps according to that level
Additional research needed to find out what inputs could give more personalization, (e.g cuisine)
Interview Highlights & Usability Session Insights:
There is room for more research to identify the right inputs for personalization.
Some users found it confusing and tedious to type out each ingredient
Navigation and UX writing needs to be improved to encourage completion of recipe generation tasks.
Design & Prototyping
We utilized no-code platforms to build within the application, allowing us to rapidly build and test the user interface and experience without extensive coding.
After an 8-hour deep dive, our team decided to use Glide and Make to utilize the use of webhooks for our MVP, to ensure seamless functionality.
Key Features
Allow users to edit their preferences
Cooking Level: 3 options: Beginner Chef, Hobby Chef, Master Chef
Diet Types: 9 options, including No Restrictions
Allergies & Other Restrictions: 11 options, including No Allergies
Excluded Ingredients: Allow users to enter in ingredients they wish to omit from recipe results.
Our previous research using ChatGPT and Perplexity gave us insight on the most common diet types in the world, as well as the most common allergies people experience.
Enter a recipe request based on what meal to make and ingredients on hand
Meal Type: Breakfast, Lunch, Dinner, Desserts, Snack
Ingredients in your Kitchen: Enter in the ingredients available to you
3 recipes generated by AI included key components such as:
Title of each recipe
Total amount of time to cook
Ingredients needed
Recipe steps
Due to limited knowledge of AI and webhooks, recipes were missing images - a key feature we would like to improve and include in the next iteration.
Results & Impact
We presented our MVP at a “Pitch Day” event, garnering valuable feedback from the users and 12 industry experts, despite not being officially launched.
Although our insights validate the product concept by addressing common pain points and measuring user satisfaction, there is always potential for further improvement:
Broaden User Research: Engage a wider audience to capture diverse insights and ensure the solution effectively addresses varied user needs.
Iterative Development: Use user feedback to refine features and enhance the overall user experience continuously.
Key Metrics Focus: Concentrate on crucial KPIs such as user acquisition, engagement, retention rates, and satisfaction to validate the product concept and drive improvement.
If we had more time…
01.
Refine the recipe instructions by introducing two distinct sections: “Ingredients on Hand” and “Ingredients Still Needed.” This will ensure users have all the necessary information at their fingertips as they cook.
02.
Enhance personalization in recipe instructions by incorporating portion sizes and detailed nutritional information, motivating users to keep cooking with the app.
03.
As accessibility is always top-of-mind for designers, we would like to explore additional strategies for making our product more user-friendly to individuals outside of our target group, particularly those who rely on screen readers and similar assistive technologies.
Reflections
This experience was unique, as I had the opportunity to collaborate with a cross-functional team, including a Product Manager and two UX Researchers. We leveraged tools such as Figma/FigJam and EdApp from the start, which played a key role in guiding our team. Although we encountered several challenges and had to interpret assignments and exercises on our own, we collaborated closely with the founders to successfully meet our deadlines.
As a designer, I am constantly pushed to explore both innovative and practical ideas. The most significant challenge was developing MVP concepts with basic features. Overcoming this challenge helped align my understanding of the ideation phase and improved collaboration with my team.
I appreciate the chance to gain practical experience with AI and a no-code tool like Glide while working in a team. This experience was distinct and unlike anything I’ve been involved in before.
Future Plans
Future developments may include conducting more extensive research and validating problems and ideas early in the program. This approach will help ensure the team builds the right solution from the start, conserving time and resources while boosting the project's overall success and impact.
As the team progresses, multiple rounds of revisions will be undertaken to improve usability and functionality. We intend to continually refine and expand the MVP based on user feedback.