Overview 01
What is UX Rival?
UX Rival is an AI-powered competitive UX analysis tool that allows designers, product managers, and founders to instantly compare the user experience of two digital products. You input your product URL, a competitor's URL, and optionally define a focus area — and the tool returns a structured, AI-generated analysis covering navigation, onboarding, visual hierarchy, feature gaps, and actionable recommendations.
It was built out of a direct personal frustration: competitive analysis is one of the most important but most time-consuming parts of the design process. Manual methods produce inconsistent, shallow insights. UX Rival makes the process fast, consistent, and exportable.
~30s
Time to full analysis
Solo
Design + engineering
100%
Vibe-coded & shipped
The Problem 02
Competitive analysis was broken.
Traditional competitive UX analysis means: open competitor's app, take screenshots, write notes in Notion, try to draw comparisons, spend 3–6 hours, produce something that's already outdated. Product designers do this constantly — and it's one of the least loved parts of the job.
The problem isn't just time. It's consistency. When the analysis is manual, it's filtered through the analyst's biases and current mood. There's no standardised framework, no reliable rubric, and the output format changes every time.
⏱
Time drain
A proper competitive UX analysis takes between 3–8 hours manually. Designers deprioritise it, skip it, or do it badly under time pressure.
📊
Inconsistent outputs
Without a standard format, insights from competitive analysis are hard to compare across projects or share with stakeholders meaningfully.
🧠
Shallow coverage
Manual analysis tends to focus on visuals and features. Deeper UX patterns — flows, friction points, trust signals — are often missed.
🔁
No reusable foundation
Each analysis starts from scratch. There's no shared intelligence across analyses, no pattern library built over time.
Design & Build Process 03
From insight to shipped product — solo.
UX Rival is a testament to the AI-augmented design workflow I've been developing. The product was ideated, designed, and engineered entirely by me — without a co-founder, without an engineering team, and without a large budget. This is what I call vibe-coding: using AI tools to compress the gap between design thinking and working code.
1
Problem validation & scoping
I validated the problem through conversations with other designers and by auditing my own past workflow. The insight was simple: everyone does competitive analysis; almost everyone hates how long it takes. I scoped the MVP tightly: two URLs + an optional focus area = one structured analysis. No accounts, no dashboards — just the core value.
2
Information architecture & prompt engineering
The most critical design decision in an AI tool isn't the interface — it's the prompt. I spent significant time engineering the system prompt that structures the AI's analysis output: what dimensions to cover, how to weight them, how to present recommendations vs. observations. This is invisible design work, but it's everything.
3
UI design & Next.js build
I designed a minimal, focused interface that doesn't get in the way of the core task. The layout guides users from input to insight without friction. Built on Next.js with a clean API integration layer — starting with Anthropic's Claude API before migrating the AI backend to Google Gemini 2.0 Flash Lite for better performance on the specific use case.
4
Feature iteration: Focus Areas, modal results, export
Post-launch, I added three key features based on usage: a Focus Areas input that lets users direct the AI's attention (e.g. "focus on onboarding" or "analyse checkout flows"), a modal-based results display for better readability, and an export function so designers can drop insights directly into reports and Notion pages.
5
Strategic positioning
Beyond the product itself, UX Rival became a portfolio centrepiece demonstrating my ability to ship real tools. I've used it in job applications, freelance pitches, and thought leadership content — it's proof of concept and portfolio signal simultaneously.
Technology 04
Built with a lean, modern stack.
The stack was chosen for speed of execution, not prestige. Every choice serves the goal of shipping a working product solo, fast.
Next.js
Google Gemini 2.0 Flash Lite
Prompt Engineering
React
Tailwind CSS
REST API
Vercel
originally Claude API
The migration from Anthropic's Claude to Gemini 2.0 Flash Lite was a deliberate product decision — the lighter model offered lower latency for the specific structured output task without sacrificing analysis quality. This kind of model-level product decision is what distinguishes a design-engineer from a pure designer.
Impact & Reflection 05
What UX Rival actually proves.
The product works — but what it proves matters more than what it does. UX Rival is evidence that I can take a product from zero to shipped without depending on an engineering team. It demonstrates prompt engineering as a design skill, and it shows that I think in systems: the input/output logic, the information architecture of the AI's response, and the export workflow are all designed, not just coded.
It's a living portfolio piece. Every iteration teaches me something about building AI tools, and the codebase has become a template I reference for new projects. The tool is actively used — by me, by peers, and by people who found it through my @techwithtomiwa content.
🎯
Proof of AI-fluent design
Building UX Rival demonstrated that I don't just use AI in my design process — I can design the AI layer itself, including prompt architecture and output structure.
🚀
Portfolio & career signal
Used directly in applications for AI Product Designer roles and UK Global Talent Visa evidence, UX Rival functions as a tangible signal of my capability at the intersection of design and AI engineering.
