AI as a nutritionist

Redesigning calorie tracking experience with AI

Date

06/2024 - 01/2025

Role

Co-founder

Download

Background

NomNow is an AI-powered calorie tracker using multimodal input (voice, image, text) and a RAG-based AI agent to make food logging fast and personalized. Built for East Asian users, it solves pain points like tedious mannual food tracking and portion size estimation.

My Role

As the co-founder, I participated in the whole process.

I mainly led the multimodal AI interaction design, crafted the product and market strategy and drove project management using agile methods.

team

One Product Manager(🙋)
Three Developers
Two UI/UX designers

Further Reading of NomNow

01. Discover

Problem observed

Have you ever tried to track your calories intake, only to give up because it was too cumbersome?

Pain points

While there are numerous traditional calorie tracking apps available, they fail to address these specific challenges.

Logging food manually is often seen as a hassle.

Users struggle to measure correct food portions.

Technical background (Untill June 2024)

Multimodal AI models
- GPT-4 Turbo enabled natural, real-time voice interaction.
- Multimodal models improved image understanding and visual input accuracy.
- AI agents and LLMs powered contextual reasoning and task automation.

Recent breakthroughs in AI open up new possibilities
for transforming the calorie-tracking experience.


But before we jump in—is it a business worth building?

02. Define

Weight loss vs. Fitness / Nutrition

Calorie tracking is about many things. What specific need are we solving?

Unmet Demand in a Proven Market

Weight loss is a persistent, proven, monetizable need, without satisfying solution.
It is an urgent, emotionally-driven needs🤯 that supports self-improvement 😀,
thus leading to more willing to pay💰 than other health goals.

We target at young, busy, especially female group

We target young people with busy lives because they care about health but struggle to form consistent habits. They value tools that are efficient, intuitive, and easy to integrate into their daily routines.

Market research insights

The global health market is vast and expanding. There's still blank space in the AI calorie tracker market.

Our team is familiar with the Greater China market, Japan, and North America. We researched these markets and got the following insights.

❌ Mainland China: While it boasts a large market size, strict AI regulations, low user payment habits, and heavy industry monopolization pose significant challenges.

❌ North America: The presence of a successful competitor, Cal AI, has validated the market’s profitability, but intense competition makes it less favorable for entry at this stage.

✅ East Asia (Hong Kong, Macau, Taiwan, and Japan): These markets remain untapped by similar AI-driven products. With strong user willingness to pay and substantial market size, they offer significant growth potential.

We have chosen to start with Hong Kong, Macau, and Taiwan due to their favorable payment habits and market readiness, followed by expansion into Japan.

Product Key Features

Standing in Hanako’s shoes, we defined our product’s key features.

👩

Personalized

AI offers tailored and expert feedback to support your
weight-loss journey.

🙌

Effortless

Quickly log meals using voice, photos, or text. AI accurately tracks your calorie.

⚖️

Accuracy

Our RAG based agent and tailored database helps accurate tracking.

SWOT Analysis

Now let's wrap up our current situation in a SWOT analysis.

03. Design

Version 1.0 (unadopted)

Iteration

Feedbacks from user test:

1. Optimize how the IP represents “percentage/calorie intake.”
2. Add more engaging interaction between the users and the IP.
3. Ensure the color and IP aligns with a food-related theme, as the current design does not resemble anything edible.

Design system

Multimodal AI Interaction

Track your calorie in a more flexible way.

Track calorie intake via Text.

Track calorie intake via Voice.

Track calorie intake via Photo.

Home page

We use IPs of different colors to visually represent today’s calorie intake status (too little, adequate, or excessive).

IP design process

New IP design (adopted)

04. Deliver

UI design

We have Chinese simplified, Chinese Traditional and English version.

User Reviews

We launched our app on app store and received 5000+ downloads.
Here's the user reviews.

Michigan Medicine