Effortlessly Track Meals to Uncover Food Intolerances with AI-Driven Insights.

I designed an end-to-end solution to reduce friction for users when they log and analyze meals. The app leverages machine learning to remove the guesswork from identifying ingredient triggers that often confound users.

Responsibilities

UX/UI Design

User Testing

User Interviews

Duration

March 2025 – 

April 2025

recommendation calendar
recommendation calendar
Homepage
Homepage
meal analysis
meal analysis
analysis discovery
analysis discovery
analysis triggering meals
analysis triggering meals
completed recording
completed recording
delete meal
delete meal

25%

of Americans reported having at least one food intolerance.

woman confused by meal
woman confused by meal
woman confused by meal

Sensitivities Are Hard to Identify

After eating it can take hours, days or even weeks for symptoms to present

Multiple foods acting together may cause the irritation

Stress, sleep patterns, salt intake and even humidity are contributing factors

Elimination strategies like the FODMAP diet can take years to succeed

I spoke with a GI Doctor and learned patients are asked to write down their meals for two weeks...
patients often quit after a few days

I wanted to understand why so...

I asked
3 people

to log meals for
3 days

Here’s what they told me:

“Entering my meals was slow and terrible.
I considered quitting but it was for only three days.”
– Jessica R.

“I don’t see myself following through if I had to enter meals for months to identify what irritates me.”
– Aaron H.

I used their insights to inform food logging

Users may choose from multiple pathways to log meals

Text entry

By voice

Food scan

Barcode scan

Pick from list

Recipe link

Meal logging options
Meal logging options
Meal logging options

Meal Verification
Interaction Exploration

Most modes of food entry will share a common design to edit, add, or remove entries. I explored multiple layouts and solutions.

  • Verify your meal version 1

    List style

    Individual item delete, bulk delete

  • Verify your meal version 2

    Grouped Card Layout

    Tap with delete and bulk delete

  • Verify your meal version 3

    1

    Individual Card Layout

    Swipe to delete

  • Verify your meal version 4

    Individual Card Layout

    Long press to delete

Verify your meal version 1
Verify your meal version 1
Verify your meal version 2
Verify your meal version 2
Verify your meal version 3
Verify your meal version 3

1

Verify your meal version 4
Verify your meal version 4

List style

Individual item delete, bulk delete

Grouped Card Layout

Tap with delete and bulk delete

Individual Card Layout

Swipe to delete

Individual Card Layout

Long press to delete

Why the interaction was chosen

1

The card layout with swipe to delete interaction method was chosen.

User testing revealed a preference for swipe interactions for ingredient deletion and subsequent interactions for meal rating.

The card layout with swipe to delete interaction method was chosen.

User testing revealed a preference for swipe interactions for ingredient deletion and subsequent interactions for meal rating.

The card layout with swipe to delete interaction method was chosen.

User testing revealed a preference for swipe interactions for ingredient deletion and subsequent interactions for meal rating.

The swipe
micro interaction

The voice logging journey

  • Foodlens homepage

    After eating dinner, Jessica opens the FoodLens app to enter her meal.

    She spots the enter meal button and taps the button to enter her meal.

  • Foodlens meal log entry page

    She scans through the entry options and taps to log her meal by voice.

  • Foodlens choose your meal

    Next, she taps dinner to log her meal.

  • Foodlens begin voice entry

    She is greeted with a screen informing her how to describe her meal. After reading the instructions, she taps begin voice entry.

  • Foodlens voice entry instructions

    She reads the prompt and begins to enter her dinner, Chicken Marsala, and then lists her ingredients. When she is done she tells the system I’m finished.

  • Foodlens verify voice entry

    When Jessica finishes speaking about her meal, she is prompted to verify the results.

    She taps verify voice entry to continue.

  • Foodlens delete an entry

    The second ingredient, two scoops of vanilla ice cream, is incorrect. She swipes left, revealing a drawer to delete the item.

    She taps the icon to remove the entry.

  • Foodlens add an ingredient

    Next she remembers she forgot to enter a key ingredient and taps the plus icon to enter a new field.

  • Foodlens type in field

    Jessica begins typing in the new field, enters Marsala wine, and taps return once complete.

  • Foodlens complete verification

    After she adds Marsala wine the remaining ingredients appear correct, she taps finish to complete her meal entry.

  • Foodlens meal entry completed

    Jessica successfully completes entering her meal.

User testing the flow revealed

Key Insight

4 out of 5 users stated they would be more likely to log meals over a long-term period with multiple modes of entry.

Insight

Users want to accomplish their task as quickly as possible. Reduce unnecessary taps.

Users want to accomplish their task as quickly as possible. Reduce unnecessary taps.

Users want to accomplish their task as quickly as possible. Reduce unnecessary taps.

meal verification success state
meal verification success state
meal verification success state
meal verification tap to approve
meal verification tap to approve
meal verification tap to approve

Solution

I removed the verification step for each field. Users understood they can can edit a field by tapping on it.

I removed the verification step for each field. Users understood they can can edit a field by tapping on it.

I removed the verification step for each field. Users understood they can can edit a field by tapping on it.

meal verification tap to edit
meal verification tap to edit
meal verification tap to edit

Insight

During testing users didn’t understand how to complete voice entry.

During testing users didn’t understand how to complete voice entry.

During testing users didn’t understand how to complete voice entry.

Solution

I improved labelling to clearly communicate how the interaction ends.

I improved labelling to clearly communicate how the interaction ends.

I improved labelling to clearly communicate how the interaction ends.

Key Learnings

User abandonment can be reduced by providing multiple paths to accomplish a goal.

User abandonment can be reduced by providing multiple paths to accomplish a goal.

User abandonment can be reduced by providing multiple paths to accomplish a goal.

Remove unnecessary steps to help users reach their goals quickly.

Remove unnecessary steps to help users reach their goals quickly.

Remove unnecessary steps to help users reach their goals quickly.

Ensure instructions are clearly defined from beginning to end to remove ambiguity.

Ensure instructions are clearly defined from beginning to end to remove ambiguity.

Ensure instructions are clearly defined from beginning to end to remove ambiguity.

Project reflections and challenges

Extend voice entry by designing a conversational interface to interact with users naturally, add shortcuts for favorites, and speech-based ingredient editing.

Extend voice entry by designing a conversational interface to interact with users naturally, add shortcuts for favorites, and speech-based ingredient editing.

Extend voice entry by designing a conversational interface to interact with users naturally, add shortcuts for favorites, and speech-based ingredient editing.

Examining user preferences proved helpful with identifying an interaction method to remove ingredients. It should be confirmed by testing with a larger sample size.

Examining user preferences proved helpful with identifying an interaction method to remove ingredients. It should be confirmed by testing with a larger sample size.

Examining user preferences proved helpful with identifying an interaction method to remove ingredients. It should be confirmed by testing with a larger sample size.

I found designing for voice interactions challenging. The breakthrough occurred when I spoke with a professor about their published paper concerning conversational user interfaces (CUIs) and food logging. The findings proved invaluable to understand user preferences.

I found designing for voice interactions challenging. The breakthrough occurred when I spoke with a professor about their published paper concerning conversational user interfaces (CUIs) and food logging. The findings proved invaluable to understand user preferences.

I found designing for voice interactions challenging. The breakthrough occurred when I spoke with a professor about their published paper concerning conversational user interfaces (CUIs) and food logging. The findings proved invaluable to understand user preferences.

© Michael Davis 2025