From Mockups to Merged PRs: An AI-Native Designer’s Playbook
Xinran Ma with Ayon · 1 November 2026
Read the originalAyon, a designer at the AI start-up Eve, edits production code with Claude Code and teaches other designers to prototype in code, so designs reach customers as intended.
Xinran Ma, who hosts interviews with designers, published a condensed interview with Ayon, the seventh employee of Eve, an AI software start-up backed by the venture firm Andreessen Horowitz. Ayon no longer uses Figma at work: prototypes are built with Claude Code, Anthropic's AI coding tool, and small visual fixes go straight into the production code. Most of Eve's designers now work in a simplified copy of the app that Ayon built for prototyping.
- Engineers usually built 75 to 80 per cent of a design, and annotated documents listing the differences never closed the gap.
- Ayon now ships small layout, spacing, colour and typography changes to the production code with AI tools, within guardrails agreed with engineers.
- Eve's designers, most without a technical background, prototype in a simplified copy of the app with fake data, use Git and deploy their work on Vercel.
- To win access to a codebase, Ayon advises building a relationship with a tech lead over several one-to-one meetings before asking.
A design team can keep a lightweight copy of its product for prototyping, so that designers test real interactions and AI behaviour without touching the production system. The interview ends with a promotion for Ayon's paid course, and the results are Ayon's own account.
Agree written guardrails with designers so that they can ship small layout, spacing and colour fixes to the production code themselves.
Keep a simplified copy of the product with fake data where designers can prototype in code without waiting on or breaking the production system.
Derived by Working Surface from the article; more in the Playbook. Source line: Now I just try to ship production code to close that last 20%.
2 November 2026: as AI tools speed up design, engineering and writing, teams name a person who must be able to explain and defend the result, because the tools' own reports and specifications cannot stand in for that person.