ompsure

Research to Website in 26 Minutes

Two Design Forge skills — UX Researcher for competitive intelligence, then Design Forge for implementation — produced a full 13-page Next.js marketing site with Radix mega menus, 5 industry verticals, and research-driven branding.

From Research to 13-Page Website in 26 Minutes

Two Design Forge skills — UX Researcher for competitive intelligence, then Design Forge for implementation — produced a full Next.js marketing site with Radix mega menus, 5 industry verticals, interactive playground, and research-driven branding. Total hands-on build time: 26 minutes.

What Happened

A developer needed a full marketing website for AIGRaaS (AI Guardrails as a Service) — a new product entering a $1.9B market with 6+ direct competitors. The workflow used two Design Forge skills across two sessions:

Session 1 — UX Researcher produced two research documents: a competitive analysis (6 competitors deep-dived, gap analysis, differentiation matrix) and a full market intelligence + design direction report (market sizing, 4 personas, messaging framework, validated pricing, complete visual direction with hex codes and font choices, 35+ sourced citations).

Session 2 — Design Forge read both research reports, loaded the taxonomy routing, and built a complete 13-page website with 47 files and 4,244 lines of production code — from create-next-app to clean build in 16 minutes, with 3 rounds of live mega menu iterations bringing the total to 26 minutes.

Phase 1: UX Researcher

The /design-forge:ux-researcher skill was invoked twice in a prior session, producing two comprehensive reports:

ReportLinesWordsContent
Competitive Analysis1,0827,2106 competitors deep-dived, gaps analysis, differentiation matrix, messaging analysis
Market Intelligence + Design Direction7206,072TAM/SAM, 4 personas, pricing validation, full visual direction, 35+ citations
Total1,80213,282Complete product positioning + website design brief

What the Researcher Discovered

Critical finding — consolidation wave: 3 of 7 primary competitors were acquired in 18 months (Lakera → Check Point ~$300M, CalypsoAI → F5 $180M, Aporia → Coralogix ~$50M). This became a key messaging angle on the final site: "Their roadmaps now belong to networking, logging, and firewall companies."

Gap analysis that drove the entire site design:

GapWho Has ItAIGRaaS Fills It
Voice AI guardrailsNobodyVAPI/11Labs/Bland integration, sub-10ms for real-time voice
Constitutional behavioral evalNobody4 decision trees, 8 harm variables, 3 validation tests
Deterministic (no LLM judge)NobodyPure TypeScript, no adversarial attacks on the judge
Operator-first no-codeAporia (now Coralogix)Dashboard-first, describe domain in English
Sub-10ms guaranteedLakera (injection only)All evaluation types, not just prompt injection

Visual direction decisions (all research-driven):

  • Teal-mint primary (#34D399) — differentiates from purple-heavy competitors, signals "safe/approved/go"
  • Dark mode default — 87% of target audience uses dark mode
  • Geist typography — matches VAPI's ecosystem (primary customer segment)
  • Animation intensity: gentle — research overrode taxonomy's "expressive" default for professional credibility

Every major design decision — color, typography, animation level, pricing tiers, page structure, CTA hierarchy — came directly from the research reports. The Design Forge build phase made zero guesses about positioning or visual direction.

Phase 2: Design Forge Build

Timeline

Time (ET)MilestoneElapsed
3:20 PMDesign Forge invoked with research reference0:00
3:22Research files read, taxonomy loaded with overrides~2m
3:22:58create-next-app scaffolded~3m
3:25Design system complete (globals.css, brand tokens, theme)~5m
3:28Shared UI + animation components written~8m
3:30Homepage with 8 sections assembled~10m
3:35All interior pages done (How It Works, Pricing, 5 Use Cases, Playground, Blog, About)~15m
3:36First clean build — all 13 routes statically generated~16m
3:40Radix refactor (NavigationMenu + Sheet replacing Framer Motion)~20m
3:46Mega menu layout iterations complete (3 rounds of live feedback)~26m

What Was Built

CategoryCount
Files created47
Lines of code4,244
Routes13 (all static)
Section components9
UI primitives7 (including Radix NavigationMenu + Sheet)
Layout components6
Animation components3
Data constants450+ lines (zero inline data)

Pages

RouteContent
/Hero + social proof + feature bento + how-it-works + stats + competitive edge + testimonials + CTA
/how-it-works5-phase constitutional pipeline, 3 evaluation modes, 3 validation tests, architecture diagram
/use-casesOverview grid → 5 verticals
/use-cases/voice-aiVAPI/11Labs/Bland integration, pain points + solutions
/use-cases/healthcareHIPAA compliance, telehealth, audit trails
/use-cases/financialPCI DSS, FINRA, hallucination losses
/use-cases/customer-serviceAir Canada ruling, unauthorized promises
/use-cases/real-estateFair Housing Act, steering language detection
/pricing4-tier cards (Free/$29/$99/Enterprise) + FAQ
/playgroundInteractive evaluation demo, 4 examples, ruleset selector
/blogArticle grid with categories, 6 research-backed posts
/aboutMission, values, timeline, Anthropic attribution

Mid-Session Refactor: Framer Motion → Radix

The initial mega menu used Framer Motion for animations. Based on the Composure Pro website's existing pattern, this was refactored to Radix UI primitives mid-session:

ComponentBeforeAfter
Desktop mega menuFramer Motion + manual stateRadix NavigationMenu with viewport={false}
Mobile navFramer Motion slide + manual backdropRadix Sheet with focus trap + portal
Nav itemsInline divs with hover handlersNavLinkCard wrapping NavigationMenuLink asChild

The Radix approach handles keyboard navigation, focus management, aria attributes, and escape dismissal — all of which the Framer Motion version missed.

How Design Forge Made This Possible

1. Research eliminated decision paralysis

Without the UX Researcher output, the developer would have needed to manually research 10+ competitors, decide on colors/typography/animation without market context, guess at pricing, and define site structure without persona analysis. The research reports provided every design decision pre-made with rationale.

2. Taxonomy routing selected the right patterns

The researcher's Classification block (saas / website / modern) fed into the Design Forge taxonomy, auto-selecting layout patterns, animation intensity, scroll choreography, and section templates — with research overrides where the researcher's findings disagreed with defaults.

3. Data-driven architecture prevented coupling

All content lives in constants.ts (450+ lines). Pages are pure assembly. Adding a use case = one object in an array + a 10-line page file. Changing navigation = edit one array, all menus update.

4. Reference site patterns matched production quality

The research identified specific patterns to borrow from Vercel (restraint, Geist typography), Stripe (mesh gradient hero), Linear (dark mode), and VAPI (color palette) — with explicit notes on what to skip from each.

Key Metrics

MetricValue
Research output1,802 lines, 96KB, 2 reports
Sources cited35+
Competitors analyzed10 (6 deep, 4 adjacent)
Build time (to first clean build)~16 minutes
Build time (including iterations)~26 minutes
Files created47
Lines of code4,244
Routes13 (all statically generated)
Iteration rounds with live feedback3

Next in the Series

This use case ends with a clean local build at 3:46 PM ET. For the deployment phase — GitHub repo creation, Shipyard preflight, Vercel deploy, live header audit, and security header hardening — continue with Hardened Deploy in 22 Minutes.

Together, the two use cases cover the full workflow from /ux-researcher invoked to aigraas.com live with A/A+ security headers in under 70 minutes of hands-on time.

On this page