top of page
Untitled_Artwork 8.gif
Device 14PM.png

Smarter Meal Logging for Patients

66%

Time on Task

8x

Task Completion

Pink Poppy Flowers

Meal logging was the most frequently used
and most abandoned feature. 

Problem

It is a daily task to be done, but it is often bounced off  by the users

The original meal flow had 7 steps.

Group 10.png

Time-on-task was high.
Drop-offs were highest here.
Users described it as “tiring” and “too much typing.”

The core issue wasn’t complexity.
It was effort.

Research Insights

From usability observation and feedback:

 

  1. Users didn’t understand serving units (grams vs household measures).

  2. Repetitive date/time selection felt redundant.

  3. Typing daily reduced motivation.

  4. Users wanted feedback, not just numeric entries.

One key behavioral insight:

If logging feels like paperwork, it won’t become a habit.

 

Pink Poppy Flowers

Real user dialogues

What I changed

I made it more visual, less texty, habit-friendly, and... we brought in AI!

This flow has something for everyone. The user can click, type, snap, or talk to AI.

#1

Quick Log 

Half of the users reportedly followed the meal plan holistically, which made it easier for us to get their logs. The user can just tick off the whole meal course in the plan. This reduces typing drastically!

Pink Poppy Flowers
custom-4.png

If the meal item isn't available

custom-5.png

Add it to Credo's database

#2

Type it Out

Since the other half of the users didn't completely follow the plan, we allowed them to enter their custom meals

planed1.png
planed2.png
custom-3.png

Select '+ Add Custom Meal'

Search & add to list

Confirm Meal course & Save!

Cam2.png

#3

or let AI do the job!

AI Food Image Recognition & Voice bot

were brought in to support quick logging!

Snap a picture & AI will do the rest

If AI fails you can always add it!

Cam1.png
Voice.png

Or talk to AI with the helper script!

Did It Work?

Yessss! People actually easily logged in their meals for all courses

66%

Time on Task

8x

Task Completion

Our AI systems had data to train the model,

Doctors were well aware of patient's meal adherence as well

bottom of page