--- title: AI agents, one pipeline: how we built a pre-sales design... description: A product designer describes a nine-agent AI pipeline built on Claude Code to accelerate pre-sales design work for RFP responses. The pipeline includes a scout... canonical: https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AI agents, one pipeline: how we built a pre-sales design studio | daily.dev og:description: A product designer describes a nine-agent AI pipeline built on Claude Code to accelerate pre-sales design work for RFP responses. The pipeline includes a scout... og:url: https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y og:image: https://api.daily.dev/og/posts/DIN6Itw1y.png og:image:alt: AI agents, one pipeline: how we built a pre-sales design studio 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. # AI agents, one pipeline: how we built a pre-sales design studio **[Medium](https://daily.dev/sources/medium_js)** · 7 min read · 1 upvotes · 0 comments ## Summary A product designer describes a nine-agent AI pipeline built on Claude Code to accelerate pre-sales design work for RFP responses. The pipeline includes a scout agent that polls SAM.gov for opportunities, a research agent that outputs to FigJam, a design system agent using Figma Console MCP to read component inventories, a UI generation agent constrained to bound design tokens only, a QA audit agent that checks for token drift, a prototype agent that outputs both clickable prototypes and code files, and a proposal agent that assembles the final pitch. Human gates sit between every stage, and the framing is 'AI enabled, not AI generated' — no output is shippable without senior designer review. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I stop an AI agent from inventing design tokens or hardcoding hex values instead of using my Figma design system? Enforce a strict no-invented-variants rule in the generation agent: values must come from bound tokens at the semantic and component level, and if a pattern isn't in the design system, the agent flags the gap instead of drawing something close. A separate QA audit agent then re-checks whether components are real and tokens are bound rather than hardcoded, catching drift before a human reviews the file. _Teams building token-safe AI design pipelines can follow ongoing coverage of agent tooling on daily.dev._ ### What does Figma Console MCP do and who built it? Figma Console MCP, built by TJ Pitre at Southleft, treats a Figma file as a queryable API by executing directly against the Figma Plugin API. This lets an AI agent traverse the full file, read variable collections, resolve alias chains, and extract component structure. It is open source, MIT licensed, and actively maintained. _Developers wiring Figma into agent workflows can track MCP tooling updates on daily.dev._ ### Why did my Figma Console MCP integration work in one context but silently fail in another? The MCP endpoint surfaces under two different tool namespaces depending on whether the environment is interactive or headless, so an agent configured for only one namespace works in that mode and fails silently in the other. The fix is carrying both namespaces in every agent file from the start of the build rather than discovering the mismatch later. _Anyone debugging MCP namespace quirks can keep up with practical fixes like this on daily.dev._ ## Similar posts on daily.dev - [Design Systems for AI Agents: How to Make Yours Agent-Ready](https://daily.dev/posts/design-systems-for-ai-agents-how-to-make-yours-agent-ready-stcywhhbg) · Headway · 0 upvotes · 0 comments - [How We Use AI to Turn Figma Designs into Production Code](https://daily.dev/posts/how-we-use-ai-to-turn-figma-designs-into-production-code-2nkdxmjjb) · monday Engineering · 60 upvotes · 5 comments - [4 new ways to go from idea to product with AI tools](https://daily.dev/posts/4-new-ways-to-go-from-idea-to-product-with-ai-tools-fdomwpsz2) · Figma · 0 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#mcp](https://daily.dev/tags/mcp), [#claude-code](https://daily.dev/tags/claude-code), [#figma](https://daily.dev/tags/figma), [#design-systems](https://daily.dev/tags/design-systems) [View this post on daily.dev](https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y) ```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":"AI agents, one pipeline: how we built a pre-sales design studio","url":"https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y"},"datePublished":"2026-07-28T19:33:32.402Z","dateModified":"2026-09-14T07:28:29.129Z","description":"A product designer describes a nine-agent AI pipeline built on Claude Code to accelerate pre-sales design work for RFP responses. The pipeline includes a scout...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/53967094842f667355d1327377b2c0ba?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/53967094842f667355d1327377b2c0ba?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Medium","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":"Medium","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/medium","url":"https://daily.dev/sources/medium_js"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,mcp,claude-code,figma,design-systems","timeRequired":"PT7M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Medium","item":"https://daily.dev/sources/medium_js"},{"@type":"ListItem","position":3,"name":"AI agents, one pipeline: how we built a pre-sales design studio"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/ai-agents-one-pipeline-how-we-built-a-pre-sales-design-studio-din6itw1y#faq","mainEntity":[{"@type":"Question","name":"How can I stop an AI agent from inventing design tokens or hardcoding hex values instead of using my Figma design system?","acceptedAnswer":{"@type":"Answer","text":"Enforce a strict no-invented-variants rule in the generation agent: values must come from bound tokens at the semantic and component level, and if a pattern isn't in the design system, the agent flags the gap instead of drawing something close. A separate QA audit agent then re-checks whether components are real and tokens are bound rather than hardcoded, catching drift before a human reviews the file. Teams building token-safe AI design pipelines can follow ongoing coverage of agent tooling on daily.dev."}},{"@type":"Question","name":"What does Figma Console MCP do and who built it?","acceptedAnswer":{"@type":"Answer","text":"Figma Console MCP, built by TJ Pitre at Southleft, treats a Figma file as a queryable API by executing directly against the Figma Plugin API. This lets an AI agent traverse the full file, read variable collections, resolve alias chains, and extract component structure. It is open source, MIT licensed, and actively maintained. Developers wiring Figma into agent workflows can track MCP tooling updates on daily.dev."}},{"@type":"Question","name":"Why did my Figma Console MCP integration work in one context but silently fail in another?","acceptedAnswer":{"@type":"Answer","text":"The MCP endpoint surfaces under two different tool namespaces depending on whether the environment is interactive or headless, so an agent configured for only one namespace works in that mode and fails silently in the other. The fix is carrying both namespaces in every agent file from the start of the build rather than discovering the mismatch later. Anyone debugging MCP namespace quirks can keep up with practical fixes like this on daily.dev."}}]} ```