The Burning Question
"DEV – HOW DO YOU WORK WITH AI? IS THERE ANY WORK ACTUALLY LEFT BESIDES PROMPTING YOUR EXACT NEEDS?
The real answer is a lot more interesting than your average hot take.
Prompting gets you off the starting line, but turning that raw output into scalable design systems, enterprise-grade UX, and products people actually trust still requires real strategy, design craft, and cross-functional alignment.
That starting prompt isn’t just hype; it’s the operational reality. Prompting gets you the raw material—a rough layout, boilerplate code, or initial user flows—but raw material doesn’t ship products.
Here is how I actually deploy AI across the lifecycle to build UX-centered products faster without sacrificing design integrity:
Instead of spending days manually organizing user interviews, support tickets, and desk research, I feed structured research inputs into an AI context window.
The AI Role: Synthesizes patterns, extracts edge-case user personas, and flags conflicting user behaviors instantly.
The Human Craft: Validating those findings against real-world human behavior, ensuring we aren't designing for synthetic assumptions, and mapping out authentic user journeys.
AI excels at spatial and visual permutations. By pairing LLM reasoning with generative UI frameworks, I skip hours of manual auto-layout setup for initial explorations.
The AI Role: Generates multi-variant wireframe structures and responsive layout options in seconds.
The Human Craft: Evaluating hierarchy, accessibility (WCAG), emotional tone, and ensuring the interface directly solves the core user problem rather than just looking good.
The biggest friction in shipping product is the gap between design tokens and production code. Using AI toolchains (like Cursor or custom script agents), I translate design system logic directly into reusable front-end code.
The AI Role: Generates accessible, typed component variations that map 1:1 to our token library.
The Human Craft: Enforcing systemic consistency, refining micro-interactions, and auditing responsive edge cases so engineering gets clean, shippable code right out of the box.
Traditional handoff documentation is often where momentum dies. AI bridges the gap between design intent and engineering execution by auto-generating clear spec notes and edge-case behavior docs directly from canvas states.
The AI Role: Writes initial unit tests, generates dynamic dummy data for dynamic state testing, and drafts technical specifications.
The Human Craft: Facilitating cross-functional alignment across QA, Product, and Engineering—ensuring the team builds with high velocity and zero guesswork.