--- title: The End of the Static Screen: Architecting Intent-Driven... description: A commercetools GM shares lessons from building an AI-driven, intent-based UX system that replaces static SaaS interfaces with dynamically generated UI. He... canonical: https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools | daily.dev og:description: A commercetools GM shares lessons from building an AI-driven, intent-based UX system that replaces static SaaS interfaces with dynamically generated UI. He... og:url: https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya og:image: https://api.daily.dev/og/posts/dKWD1TsYa.png og:image:alt: The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools **[AI Engineer](https://daily.dev/sources/aidotengineer)** · 23 min read · 0 upvotes · 0 comments ## Summary A commercetools GM shares lessons from building an AI-driven, intent-based UX system that replaces static SaaS interfaces with dynamically generated UI. He walks through failed early attempts where the same query produced inconsistent layouts, then explains the current architecture: an orchestrator extracts user intent, invokes first- and third-party tools/agents, and a UX agent renders components from a curated catalog. He contrasts three generative UI approaches (controlled component rendering, fully open-ended LLM-generated HTML, and a middle 'declarative' path using protocols like Google's AUI, Vercel's JSON render, and OpenUI) and explains why his team chose the declarative approach for design-system compliance. He closes with three major challenges: information architecture (arranging AI-picked components using atomic design and a layout/slot/sub-slot hierarchy), maintaining a rigorous component catalog/schema as the 'contract' between agent and UI, and the changing nature of design/PM work when pixels are no longer manually designed. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the three main approaches to generative UI rendering with LLMs? They are the controlled approach (the agent picks a predefined, opinionated component and renders it exactly as designed, as ChatGPT does for restaurant search), the open-ended approach (full autonomy is given to the LLM to generate raw HTML rendered in a sandboxed iframe, as Claude does for diagrams), and the declarative approach (a middle ground using protocols like Google's AUI, Vercel's JSON render, or OpenUI, where the LLM selects and arranges components from a defined catalog while staying compliant with a design system). _Teams weighing generative UI architectures can track protocol comparisons like these on daily.dev._ ### Why did early attempts at generative UI produce inconsistent dashboards for the same query? Giving an LLM full autonomy to compose an experience from available components without a governing layout structure caused the same query, phrased differently, to generate four different layouts with inconsistent data (one showing Q1 sales, another showing January-March) and inconsistent visual composition, making the experience confusing rather than personalized. Codifying UX knowledge into an atomic-design-based layout hierarchy of layout, slots, sub-slots, and components fixed this. _Developers building agentic UI can follow real-world debugging patterns like this via daily.dev._ ## Similar posts on daily.dev - [Generative UI in 2026: A Developer Guide That Ships](https://daily.dev/posts/generative-ui-in-2026-a-developer-guide-that-ships-didpe5oji) · Alex CloudStar · 0 upvotes · 0 comments - [A better approach to generative UI](https://daily.dev/posts/a-better-approach-to-generative-ui-hwpqccf9w) · InfoWorld · 0 upvotes · 0 comments --- Tags: [#ux](https://daily.dev/tags/ux), [#mcp](https://daily.dev/tags/mcp) [View this post on daily.dev](https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools","url":"https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya"},"datePublished":"2026-09-01T20:23:47.520Z","dateModified":"2026-09-01T20:24:11.020Z","description":"A commercetools GM shares lessons from building an AI-driven, intent-based UX system that replaces static SaaS interfaces with dynamically generated UI. He...","image":"https://i.ytimg.com/vi/QrMcNe2jjt8/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/QrMcNe2jjt8/sddefault.jpg","isAccessibleForFree":true,"articleSection":"AI Engineer","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"AI Engineer","logo":"https://media.daily.dev/image/upload/s--u5PucxNT--/f_auto/v1724338940/logos/aidotengineer","url":"https://daily.dev/sources/aidotengineer"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ux,mcp","timeRequired":"PT23M","video":{"@type":"VideoObject","name":"The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools","description":"A commercetools GM shares lessons from building an AI-driven, intent-based UX system that replaces static SaaS interfaces with dynamically generated UI. He...","thumbnailUrl":"https://i.ytimg.com/vi/QrMcNe2jjt8/sddefault.jpg","uploadDate":"2026-09-01T20:23:47.520Z","duration":"PT23M","url":"https://api.daily.dev/r/dKWD1TsYa","embedUrl":"https://www.youtube.com/embed/QrMcNe2jjt8"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"AI Engineer","item":"https://daily.dev/sources/aidotengineer"},{"@type":"ListItem","position":3,"name":"The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetools"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-end-of-the-static-screen-architecting-intent-driven-ux-gus-iwanaga-commercetools-dkwd1tsya#faq","mainEntity":[{"@type":"Question","name":"What are the three main approaches to generative UI rendering with LLMs?","acceptedAnswer":{"@type":"Answer","text":"They are the controlled approach (the agent picks a predefined, opinionated component and renders it exactly as designed, as ChatGPT does for restaurant search), the open-ended approach (full autonomy is given to the LLM to generate raw HTML rendered in a sandboxed iframe, as Claude does for diagrams), and the declarative approach (a middle ground using protocols like Google's AUI, Vercel's JSON render, or OpenUI, where the LLM selects and arranges components from a defined catalog while staying compliant with a design system). Teams weighing generative UI architectures can track protocol comparisons like these on daily.dev."}},{"@type":"Question","name":"Why did early attempts at generative UI produce inconsistent dashboards for the same query?","acceptedAnswer":{"@type":"Answer","text":"Giving an LLM full autonomy to compose an experience from available components without a governing layout structure caused the same query, phrased differently, to generate four different layouts with inconsistent data (one showing Q1 sales, another showing January-March) and inconsistent visual composition, making the experience confusing rather than personalized. Codifying UX knowledge into an atomic-design-based layout hierarchy of layout, slots, sub-slots, and components fixed this. Developers building agentic UI can follow real-world debugging patterns like this via daily.dev."}}]} ```