AI agent & app designer · Vibe coding
Vibe coding is the new production line. UX is what makes it ship.
I build and launch AI-powered apps end to end: native iOS and Android, full-stack web platforms, and the agents that automate the business behind them. Nineteen years of UX strategy, applied at the speed AI now allows.
0 · UX in the era of vibe coding
AI can write the product. It cannot feel what the product does to a person.
Vibe coding removed the build constraint. A founder can now describe an app on Monday and have a working App Store build by Friday. What it did not remove is the human constraint: attention, trust, confusion, fear, motivation, habit. Those still behave exactly the way they did before the models arrived.
So the bottleneck moved. It used to be engineering capacity. Now it is judgment. When every competitor can generate the same feature set in a weekend, the experience is the only durable difference, and experience quality is a research and empathy problem, not a generation problem.
Where UX shows up in ROI, stage by stage.
UX is not a cost line, it is a multiplier on every other line. The same traffic, the same ad spend and the same feature set produce very different revenue depending on how well the experience is designed. These effects compound, which is why well-designed products do not grow a little faster, they grow on a different curve.
Discovery
Clarity of the promise
If a visitor cannot say what your product does in one sentence, nothing downstream matters. Positioning, hierarchy and proof are UX decisions long before they are marketing decisions.
First run
Activation
The gap between install and first real value is where most AI-built products die. Reduce steps, defer signup, show the outcome early, and activation moves in double digits.
Habit
Retention
Retention is designed, not marketed. Reasons to return, respectful notifications, remembered state and progress that feels earned are what keep a cohort alive past week four.
Money
Conversion and paywall
Pricing pages and paywalls are empathy tests. Timing, framing, perceived fairness and the absence of dark patterns decide whether someone upgrades or uninstalls.
Cost
Support and churn
Every confusing screen becomes a support ticket, a refund, or a one-star review. Good UX is the cheapest customer-support budget you will ever spend.
Speed
Rework avoided
AI makes it trivial to build the wrong thing very fast. Research and prototyping before generation is what prevents three sprints of confident, well-typed waste.
Compounding matters more than any single number: lift discovery, activation, conversion and retention even modestly and the effects multiply through the funnel, while the support and rework costs fall at the same time. That is the honest case for UX spend, and it is the reason a product with strong UX outgrows a technically identical one.
Six things AI still cannot do in UX.
I use these tools every day and I am extremely advanced in them. That is exactly why I am specific about their limits. AI does not understand human emotion, it reproduces the surface of products that once responded to it.
AI has no emotional stake
A model has never felt the anxiety of entering payment details on a page it does not trust, or the shame of failing a form. It predicts what interfaces usually look like. It cannot feel what this interface does to a person at 11pm on a bad day.
It averages, it does not empathize
Generative models regress to the mean of their training data. That produces competent, familiar, forgettable products. Differentiation comes from a specific human insight about a specific audience, which is exactly what averaging destroys.
It cannot sit with a real user
There is no substitute for watching someone hesitate, tap the wrong thing, and quietly give up. That hesitation is the data. AI cannot observe it, and it cannot infer the context, culture, disability, stress or motivation behind it.
It has no accountability
A model will happily ship a dark pattern, an inaccessible contrast ratio, or a flow that misleads. Ethics, accessibility and consequence are judgment calls that a person has to own.
It optimizes the screen, not the journey
Ask for a screen and you get a good screen. Products fail in the seams: the empty state, the error, the offline case, the re-entry, the cancellation. Those are systems thinking, not prompt output.
It cannot decide what not to build
The highest-leverage design act is removal. AI is optimized to produce more. A strategist's job is to say no, and to protect the one flow that actually creates value.
The practical answer is not to reject AI, it is to put a designer in front of it. Research first, decide what deserves to exist, then let the agent build at full speed against a decision that was made by a human who understands the person on the other side of the screen.
1 · Native iPhone and Android AI app builders
Platforms I am extremely advanced in.
These matter when you want an actual App Store or Google Play app, not just a mobile website. The column that decides a product's future is code ownership: it is the difference between a prototype you can grow and a demo you have to rebuild.
| Platform | Output & ownership | My take |
|---|---|---|
| Newly (formerly Natively) | React Native + Expo, full source ownership, App Store and Play compliance | ★★★★★ |
| Expo Agent | Expo + React Native, agent-built native projects you own | ★★★★★ |
| Rork | React Native / Expo output | ★★★★★ |
| Rork Max | Swift and native Apple output | ★★★★★ |
| RapidNative | React Native + Expo | ★★★★★ |
| Replit Agent | Expo / React Native plus web | ★★★★★ |
| Dreamflow | Flutter, source ownership | ★★★★ |
| FlutterFlow | Flutter, visual plus code export | ★★★★ |
| Draftbit | React Native | ★★★★ |
| Vibecode | AI-generated mobile apps | ★★★★ |
| Anything | AI mobile apps | ★★★★ |
| Adalo | Proprietary native platform, no code export | ★★★ |
| Thunkable | Native mobile, no code export | ★★★ |
| Bravo Studio | Figma to mobile, partial ownership | ★★★ |
| Appy Pie | No-code mobile | ★★ |
Note: Natively was renamed to Newly in March 2026. It generates React Native + Expo projects, supports iOS and Android, provides source-code ownership, and now emphasizes App Store and Play Store compliance.
2 · Expo deserves its own category now
Expo is no longer just the infrastructure under React Native.
Expo Agent, currently in beta, lets you describe a native app and have the agent build a real Expo project. You own it, download it, push it to GitHub, work locally, and ship on Expo's infrastructure. That ecosystem is becoming essential for serious AI-built mobile apps.
- Expo Agent
- Expo Skills
- Expo MCP
- Expo Go
- EAS Build
- EAS Submit
- EAS Update
- EAS Hosting
- Codex + Expo
- Claude Code + Expo
- Cursor + Expo
- Argent
- agent-device
3 · AI full-stack and web app vibe coding
For SaaS, dashboards, portals, marketing sites and PWAs.
Lovable
Full-stack React web apps
Bolt.new
Fast full-stack web development
Replit Agent
Full-stack web plus increasingly native mobile
v0 by Vercel
React and Next.js applications
Base44
Prompt to complete application
GitHub Spark
Full-stack TypeScript/React apps with hosting, compute, AI inference and storage
Google AI Studio
AI-powered apps
Firebase Studio
Existing Firebase work; new App Prototyping Agent workspaces closed June 22, 2026
Emergent
Full-stack autonomous building
Blink.new
Full-stack web apps
Rocket.new
Prompt-based development
Create.xyz
AI application builder
Databutton
AI and full-stack apps
Softgen
AI web app generation
Same.new
Prompt and screenshot driven web building
Tempo Labs
AI development plus visual design
Onlook
Visual React development
Dyad
Local and open AI app builder
Bubble AI
AI on an established no-code ecosystem
WeWeb
Visual frontend and full-stack integrations
Softr
Business apps and portals
Glide
Data-driven business apps
Retool
Internal applications
Webflow AI
Sites and web experiences
Framer AI
Websites
Wix AI / Wix Studio
Websites and business applications
Hostinger Horizons
Prompt to web application
Durable
Business websites
One current change worth planning around: Firebase Studio stopped allowing new App Prototyping Agent workspaces on June 22, 2026, and Google is directing that workflow toward Google AI Studio.
4 · Serious coding agents
When the app graduates from prototype to production codebase.
OpenAI Codex is now a full agentic engineering environment rather than autocomplete. It works across codebases, runs refactors, migrations and features, calls tools, and coordinates multiple agents.
- OpenAI Codex
- Claude Code
- Cursor
- Windsurf
- GitHub Copilot / Coding Agent
- Google Antigravity
- Kiro
- Devin
- Augment Code
- Cline
- Roo Code
- Aider
- OpenCode
- OpenHands
- Amp
- Warp
- Zed
- TRAE
- JetBrains Junie
- Continue
- Amazon Q Developer
- Tabnine
- Google Jules
5 · The underlying stack you will keep seeing
Most of the competition is a different front end on similar infrastructure.
Mobile
- React Native
- Expo
- Flutter
- Swift / SwiftUI
- Kotlin / Jetpack Compose
Backend
- Supabase
- Firebase
- Convex
- Xano
- Appwrite
- PocketBase
- Neon
- PlanetScale
Authentication
- Supabase Auth
- Firebase Auth
- Clerk
- Auth0
Payments
- Stripe
- RevenueCat
Hosting and deployment
- Vercel
- Netlify
- Cloudflare
- Railway
- Render
- Fly.io
- Expo EAS
Code and source
- GitHub
- GitLab
AI agent + app designer
From idea to design to development to launch to growth.
Are you looking to automate your business with AI?
Are you looking to build and launch your own app?
Already have an app, but need help making it grow or go viral?
I help turn an idea, a business, or an existing product into a smarter, scalable digital platform. Build smarter. Automate more. Scale faster.
- AI Agents & Business Automation
- AI-Powered Apps & Platforms
- iOS & Android App Design
- Vibe Coding & Rapid Prototyping
- UX/UI Design & Conversion Optimization
- App Store Launch Strategy
- App Growth & Viral Marketing
- Subscription & Monetization Strategy
- Landing Pages & Sales Funnels
- Branding & Product Positioning
- AI Integrations & Workflow Automation
- Existing App Redesign & Optimization
AI marketing + growth
Visibility in AI search, not only Google.
- LLM Marketing
- AI Search Optimization
- Generative Engine Optimization (GEO)
- Answer Engine Optimization (AEO)
- LLM Optimization (LLMO)
- AI Agent Recommendation Optimization
- ChatGPT Visibility Optimization
- Google Gemini Visibility
- Perplexity Search Optimization
- Google AI Overview Optimization
- AI Brand Discovery Strategy
- AI Citation & Mention Strategy
- AI Product Recommendation Strategy
- Agentic Commerce Optimization
- AI-Readable Product & Service Data
- Structured Data & Schema Optimization
- Brand Entity Optimization
- Knowledge Graph Optimization
- Semantic SEO
- Advanced Technical SEO
- Topical Authority Development
- Content Authority Strategy
- Search Intent Optimization
- Zero-Click Search Optimization
- Featured Answer Optimization
- Digital PR & Authority Building
- Backlink Strategy
- Competitor Search Intelligence
- Local SEO
- E-Commerce SEO
- Programmatic SEO
- Conversion Rate Optimization (CRO)
App marketing + growth
Installs, activation, retention and revenue.
- App Marketing Strategy
- App Store Optimization (ASO)
- Apple App Store Growth
- Google Play Growth
- App Keyword Optimization
- App Store Conversion Optimization
- App Screenshot & Preview Optimization
- App Listing Copywriting
- App Review & Rating Strategy
- Organic App Discovery
- Paid User Acquisition
- Meta Ads
- Google App Campaigns
- TikTok App Marketing
- Instagram Growth
- Influencer & Creator Campaigns
- Viral App Growth Strategy
- Referral Systems
- Invite & Reward Loops
- Product-Led Growth
- App Onboarding Optimization
- Activation Optimization
- User Retention Strategy
- Push Notification Strategy
- Subscription Conversion Optimization
- Paywall Optimization
- App Monetization Strategy
- Lifetime Value Optimization
- Churn Reduction
- Re-Engagement Campaigns
- App Analytics & Funnel Optimization
Make your brand recommendable by AI
Your customers are no longer only searching Google.
They are asking: "What is the best app for this?"
"Which company should I use?" "What is the best product for me?"
"Which brand do you recommend?"
I position a company so AI systems can discover it, understand it, trust it, cite it, compare it, and recommend it.
- Build machine-readable brand authority
- Establish clear brand and product entities
- Create AI-citable authoritative content
- Optimize product and service information for AI retrieval
- Strengthen third-party brand mentions
- Build comparison and alternative pages
- Create authoritative FAQs and knowledge hubs
- Implement structured data and schema
- Strengthen reviews and reputation signals
- Improve consistency across the web
- Build topic and category authority
- Optimize content for conversational queries
- Track AI mentions, citations and recommendations
- Analyze how ChatGPT, Gemini, Perplexity and other AI platforms describe your brand
- Identify prompts where competitors are recommended instead of you
- Build a strategy to increase your share of AI-generated recommendations
Do not just rank on Google. Become the brand that Google, ChatGPT, Gemini, Perplexity and the next generation of AI agents understand and recommend.
Frequently asked questions
- Why does UX matter more in the era of vibe coding?
- Because AI removed the build constraint, not the human one. When anyone can generate a working app in a weekend, the code stops being the moat and the experience becomes the moat. UX is now the only layer that decides whether a product is understood, trusted, adopted and paid for.
- How does UX affect ROI and growth rate?
- UX compounds at every stage of the funnel. Clearer positioning lifts install and signup rates, better onboarding lifts activation, reduced friction lifts paywall conversion, and trust and clarity reduce churn and support load. Each of those is multiplicative, so a product with strong UX grows faster on the same traffic and the same ad spend.
- What is vibe coding?
- Vibe coding is building software by describing intent to an AI agent instead of hand-writing every line. The agent generates and iterates on a real codebase. It collapses the distance between idea and working product, which means design judgment, not typing speed, becomes the bottleneck.
- Which platform should I use to build a native iPhone and Android app with AI?
- If you want a real App Store or Google Play app rather than a mobile website, use a builder that outputs React Native + Expo or Flutter and gives you the source code: Newly (formerly Natively), Expo Agent, Rork, RapidNative or Replit Agent. Code ownership matters more than speed, because it is what lets the app survive past the prototype.
- Can AI replace a UX designer?
- No. AI generates patterns from what already exists. It does not sit with a confused user, feel the hesitation before a wrong tap, or weigh the emotional cost of a broken flow. When everyone can generate a working app, experience quality is the only remaining differentiator, and that requires empathy, context and human judgment.
- What is Generative Engine Optimization (GEO)?
- GEO, along with AEO and LLMO, is the practice of making a brand discoverable, understandable, trustworthy and citable to AI answer engines such as ChatGPT, Google AI Overviews, Gemini and Perplexity. It combines entity clarity, structured data, authoritative content and third-party mentions so AI systems recommend you rather than a competitor.
- Do you work on existing apps or only new builds?
- Both. I take ideas from concept through design, development, launch and growth, and I also redesign, optimize and scale products that are already live.
Have an idea, an app, or a business to automate?
Tell me what you are building. I answer personally.