Personal Project
Experimented with designing an app to explore how AI can support the design workflow
Personal Project | 2026
Summary
I started a personal project to experiment with AI tools and explore how UX designers can work more efficiently and effectively with AI. I decided to design an app to log a dogs you encounter in the walk.
Background
Having designed AI-powered tools at Bodygram and Indeed, I have a strong interest in the space. Today, AI is more capable and more accessible within design workflows than ever before. Through this personal project, I wanted to experiment with these tools and prepare for an AI-enabled design workflow — starting by designing a new app from scratch.
As a UX designer, I'm still figuring out how AI can be used across the entire design process. Companies want designers to use AI to be more efficient, but there's no clear playbook for how to do it effectively.
There is no established workflow or go-to guide for UX designers to use AI efficiently.
Designers are often too busy with immediate work to find time to learn and experiment.
Companies are hesitant to invest if they can't see a clear impact.
Despite this, roughly 56% of companies actively encourage AI use, while actual organizational integration for efficiency sits around 38%. (Source: Owl Labs, Gallup, Qualtrics, Worklytics, Second Talent)
56%
Roughly 56% of companies actively encourage AI use
38%
Actual organizational integration for efficiency purposes sits around 38%
Problems
I enjoy going for walks, and one of the best moments is seeing dogs out for a walk too. I track the breeds I see in my journal, but there is no way to see the summary and progress. It would be even more fun if I could log them in an app and discover new insights.
Preparation: AI Tools Research
AI used: Miro, Claude
I asked Claude and Miro what AI tools UX designers can use. I decide to continue using Claude first as it seems useful in many phases. There's a question that Claude says it is more capable than other tools, so I decided to use other tools accordingly when I need more idea.
Users
AI used: Claude
I entered what the app is about and iterated, then Claude summarized user profile as following:
Who they are
Someone who doesn't own a dog yet but is seriously thinking about getting one. They live alone or in a small household in an urban area, likely Tokyo, and regularly go on walks where encountering dogs is part of the pleasure. They're drawn to specific breeds — not just "any cute dog" — and find themselves mentally noting which ones they react to.
How they behave
They notice dogs on the street and sometimes look up the breed when they get home. They're the type of person who enjoys a bit of light research and record-keeping as a hobby in itself — not obsessively, but thoughtfully. They'd describe logging a dog encounter as a small, satisfying ritual at the end of a walk.
What they want from the app
Over time, they want the app to reflect back something meaningful — "you keep reacting to calm, elegant dogs" — that helps them make a more confident decision about what breed to eventually get. The diary aspect is a bonus: a personal record of small joyful moments.
What they don't want
Friction. They're logging after they get home, from memory, so the input needs to be fast and intuitive. They also care about doing the right thing in public — hence no camera, no capturing strangers.
Competitive Analysis
AI used: Claude
I looked at App Store to find competitors. I didn't find any but childcare app had similar concept to log the status of a child. Claude found that there's no app that exactly matches the "dog encounter logging" concept. Here are the closest ones:
BarkHappy — the most similar in spirit. It's a location-based social app where you can find nearby dogs for playdates and discover dog-friendly spots on a map.
DogLog — focused on logging your own dog's daily activities with photos, shareable with multiple caregivers. More of a care tracker than an encounter log.
DogNote — a shared pet journal for families, logging walks and activities in a real-time feed.
Ideation
AI used: Claude
According to Claude, the app has three main tabs, each mapped to one of the core goals we identified:
Diary feed — your chronological log of encounters, with individual entry cards
Breed library — every breed you've met, with a detail page and insights over time
Walk map — your encounters plotted geographically, filterable by breed
Ideation 2
AI used: Figma Make
I entered a simple prompt to generate a design — "I would like a dog encounter logging app…" — and it came up with an interactive prototype. Compared to the Claude version, it was very simple. Then I attached the screenshot from Claude for Figma Make to incorporate into the design. It updated the prototype. I wanted to tweak a few things, but it turned out that Figma Make only generates code and does not allow editing the design in Figma. So I decided to move to Figma to generate a design file so it's more flexible. Figma Make has been useful for coming up with a quick prototype, though.
Ideation 2.5
AI used: Figma AI
In Figma, I used “First Draft” to come up with the mock. I entered brief description of the app, core user profile, and tabs to include. It automatically created a social app with photos. I didn't expect the walk map and weather forecast, but they are good ideas to consider. I iterated by adding more prompts, but some prompts didn't work and timed out after a few seconds. First Draft doesn't work well for iteration, so I decided to create the design myself.
Design
AI used: Claude
I used a few plugins — Google Material Design Icons and Unsplash — for icons and placeholder images. During the design iterations, I used Claude to critique the design and explore ideas:
Can you suggest a color palette for the app?
Can you show me examples of other tabs?
Which design option would work better?
Where should the CTA be placed?
Then I finalized the design.
Prototyping
AI Used: Claude
I manually created a prototype in Figma with a simple interaction.
Reflections
Overall, it has been a fun and interesting experience working with AI in UX design. As for the app itself, I enjoyed the process and plan to iterate on the concept further. One thing I noticed is that I paused to think less than usual — instead of sitting with a problem, I turned to AI for quick opinions and critique. Whether that is always the right approach is an open question, but it provided useful knowledge and fast feedback.
Designers need to articulate their thoughts in words to use AI effectively.
There is still a gap between organic sketches (or Figma) and prompts for AI.
How might visually-oriented people — assuming designers are — translate their thoughts into words?
AI generally saves time when searching for competitors, building a basic understanding of users, and synthesizing user interviews.
There are risks such as hallucination, copyright violation, bias, and echo chambers.
The quality of image AI can create varies.
AI is good at getting started, but some tools are less effective at iteration.
Copyright©2026. Mikako Matsunaga