AI-Powered Recipe Customization
A case study on increasing menu variety and user satisfaction by allowing customers to adapt recipes to their dietary needs.
Project Overview
This project was born from the need to better serve users with specific dietary preferences (vegan, vegetarian, organic) who felt limited by the standard weekly menu. The goal was to increase the perception of variety by allowing users to customize recipes with simple ingredient swaps, leveraging AI to generate adapted recipe images and text, without drastically increasing operational complexity.
My Role
Lead Product Designer
Timeline
2 Months
Tools
Figma, UserTesting, AI Prototyping
The Challenge
Cancellation survey data was clear: nearly 15% of users who cancelled in the last year cited a perceived lack of variety as a key reason. My primary challenge as design lead was to address this by increasing options for users with dietary restrictions. However, we had to do this without overwhelming the weekly menu, as flooding the UI with too many choices would create a new pain point for our broader user base. The solution needed to be elegant, discoverable for those who needed it, but unobtrusive for those who didn't.
The Design Solution
I designed a model where customization options were discoverable directly within a recipe card. This was achieved through a subtle "nudge" system that signaled availability without causing information overload. We tested the perception of AI-generated recipe images to ensure they met our quality standards. I also designed a smart filtering system that would automatically apply customizations based on user preferences and clearly announce the change, reassuring users that the recipe was adapted specifically for them while maintaining clarity between the original and customized versions.
Customization Stack
-
Easy to track same meals with different options
-
Reduces the amount of cognitive load
Different items
-
Not hiding anything from the customer behind an action
-
It can get very overwhelming to have so many items a week
The Experience
The final end-to-end experience allows users to seamlessly discover and apply customizations. The UI nudges are integrated into the main menu, and the customization process itself is simple and intuitive, confirming the changes with new AI-generated visuals and text. This creates a powerful feeling of a personalized menu tailored to individual needs.
What difference was made?
Retention Uplift
8-Point Drop
In cancellations due to "lack of variety", measured 2 months after release.
New Revenue Stream
+$2.4 million
Generated from the uptake of customized recipes with upselling value.
Satisfaction Uplift
3.2 → 3.9 / 5
Increase in customer satisfaction score for users with dietary restrictions.
