Transforming microbiome data into AI-powered mental health and

nutritional insights

  • The Project

    Mobile App Design

  • My Role

    Product Designer

  • Timeline

    Sep - Nov 2025

  • Tools

    Figma, Notion, Lovable, Claude, ChatGPT

  • Status

    MVP Live

A Gut Feeling

Bloom is an AI-driven app that helps users understand the connection between their microbiome, gut health and mental health. By combining stool sample data, food and mood tracking, and scientific insights, this app empowers people to take charge of their gut health and emotional well-being.

View Live Prototype

Problem Statement


In my brief experience as a medical assistant in psychiatry, I noticed how patients often reported food intake symptoms alongside their mental health symptoms. The doctor I worked with often made dietary recommendations, revealing the true holistic nature of mental and physical health. Patients who were battling anxiety, depression or ADHD, often reported issues with eating less, bloating, and other gut symptoms. SSRIs specifically are well-known to contribute to weight gain, changed diets, and sleep issues - and it’s no surprise because up to 95% of serotonin is produced in the gut!

This sparked my curiosity and I picked up Dr. Emeren Mayer’s book, “The Mind-Gut Connection”. An enlightening read, it discusses the impact of diet, stress, and mindfulness practices on gut health. It also details how the gut microbiome impacts brain function and emotional regulation, specifically how dysbiosis can contribute to chronic pain, anxiety, depression and ADHD.

Most people are not aware of the connection between their gut symptoms and emotional well-being. Meanwhile, microbiome testing feels complicated, expensive, or irrelevant, and once results are obtained, they’re often confusing and underutilized. There’s a gap in tools that not only interpret microbiome data but also motivate people to begin the journey.

Where Bloom Comes in…

My mission is to demystify gut testing, visualize their gut health, and turn scientific insight into achievable, and personalized daily recommendations with a digital experience that empowers users to discover, test, and nurture their gut health. The product should:

  • reduce barriers to testing and understanding testing

  • translate results into intuitive visual feedback

  • build sustained motivation through actionable, empathetic insights for nutrition, probiotics, and mental health practices

Using AI to build an AI tool….

The design industry is evolving rapidly with the rise of AI tools. I wanted to experiment with them to accelerate my workflow and inspire new ways of designing.

I used Lovable to quickly vibe code a functional prototype for an MVP, iterating quickly within the chatbox before exporting the designs to Figma for detailed UX refinement. This process taught me not only how to prompt-engineer effectively and streamline my design flow, but also where the human element truly matters in creating human-centered design.

Define and Research Phase

To ensure the app was to built on sound research and avoids scientific errors, along with investigate the problem, I conducted some background research (with the help of Chatgpt and Claude to speed the workflow).

Heeding caution with these LLMs, I ensure that each claim came directly from the original source. I then translated the research into some possible takeaways to justify the product’s creation. 

Works Cited
Cryan, J. F., O’Riordan, K. J., Cowan, C. S. M., Sandhu, K. V., Bastiaanssen, T. F. S., Boehme, M., … Dinan, T. G. (2019). The microbiota-gut-brain axis. Physiological Reviews, 99(4), 1877–2013. https://doi.org/10.1152/physrev.00018.2018
Alli, S. R., Gorbovskaya, I., Liu, J. C. W., Kolla, N. J., Brown, L., & Müller, D. J. (2022). The gut microbiome in depression and potential benefit of prebiotics, probiotics and synbiotics: A systematic review of clinical trials and observational studies. International Journal of Molecular Sciences, 23(9), 4494. https://doi.org/10.3390/ijms23094494
Shaikh, A., et al. (2023). Irritable Bowel Syndrome and the Gut Microbiome. Journal of Clinical Medicine, 12(7), 2558. https://doi.org/10.3390/jcm12072558
Drago, L. (2025). Navigating microbiome variability: Implications for research diagnostics and direct-to-consumer testing. Microbiome.

Competitive Analysis

Most microbiome brands focus on lab analysis or diet programs, leaving out the mental health component entirely. They also tend to take a one-size-fits-all approach to the microbiome, when in reality everyone has a unique gut flora.

Conducting this competitive analysis helped me figure out how bloom will differentiate:

  • user-driven daily tracker and log as the database

  • AI and evidence-backed personalized insight

  • mental health focus (often left out of the conversation)

User Personas

I moved onto storyboarding to empathize with our target users. To keep the scientific scope manageable for an MVP, I focused on relatively healthy users — one with mild IBS and one who is a wellness seeker.

Ideation Phase

Moving into ideation, I created a few “how might we“ questions to guide possible user flows and move from abstract goals to brainstorming some potential solutions.

Pause… More Research Needed

Before I continued to sketching wireframes, I realized I still had no idea how I would integrate the microbiome test into the app. I hit pause and began doing more research. This took a while, and although I’m not a medical professional, I enjoy putting on my analytical hat and learning about topics like this one.

Initially, I thought that a stool sample might provide insights about common strains in your gut. While this is true, I’ve also learned that there are billions of gut flora. Tracking each one would be unrealistic and a scientific nightmare.

This research provided me with scientifically sound metrics of the gut microbiome, which I translated into product features and visualization ideas.

Site Map

Lo-Fidelity Wireframing

I jumped into mid-low fidelity wireframing, with what I felt would be intuitive for a home dashboard, daily log, insights page, profile page, and stool sample page.

Hi-Fidelity Prototype

I fed my lo-fi frames into Lovable screen by screen, iterating within the chatbox with detailed prompt engineering. I gave it my color palette, some UX writing based on my research findings, and the site map to design the navigation.


The best part about this is that there is no time lost to hi-fi mockups, allowing me to easily visualize and iterate.

Some downsides are the potential for blindspots, like working downstream from an unintuitive design because it has already been developed. This is why iterating with feedback and user research is imperative.

Insights Pages

The initial design that Lovable spat out was lacking hierarchy, full of formatting flaws, and missing a clear “at a glance” overview of recommendations, potentially overwhelming the user. Here are the before and after as I iterated with the chatbot and added a progress snapshot:

After (v2)

After (v2)

Before (v1)

Gut Garden

UX decisions:

  • popup text boxes to make garden free of text while allowing user to quickly gain info

  • strains classified by either relative abundance or presence, as necessary based on scientific evidence

  • key metrics shown with popup boxes to explain scientific jargon

Profile and Health Data

In designing a health app, it is imperative that users know how their sensitive data is being handled to instill trust in the app.

  • sync with Apple Health button to allow users to decide their privacy setting and how their data is shared across apps, while allowing for deeper insights

  • generate a health report PDF to share with their doctor, allowing users to inform their gut-mind insights in the app with those of real clinical professionals.

  • Learn More section to ease users’ confusions about how the app makes science-backed recommendations, onboarding info and more educational info.

Personalized Goal Setting

To ensure that the app insights can be translated into actionable changes for users, I created a goal tracking page featuring:

  • personalized goal recommendations directly from insights page to motivate users

  • rewarding messages for task completion for a little gamification

  • snapshot on home page of daily goals

  • quick edits on adding and removing goals

Daily Log

The original daily log I made in Lovable wasn’t very user-friendly. I iterated to make the log entry flow more intuitive, seamless and engaging, with minimal scrolling.

I also realized that there might be confounding between correlating a certain meal with multiple food items with a specific mood. To solve this, I incorporated a way to categorize meal type to try to stratify the food type further.

Next Steps

My next steps will be to transport the Lovable prototype back into Figma and make any UX changes that seem apparent to me. I plan on reorganizing some of the information architecture, making aesthetic/visual changes, and improving navigation through the app.

After that, I will conduct user testing before iterating again. I hope to see this app one day come to life as it becomes more scientifically feasible to use the microbiome data to predict gut-mind correlations.

Check out my other projects!