DaisyAI Website Refactor Specification v2
Purpose: Complete specification for refactoring daisyai.ai from SaaS-focused to FDE-first positioning. This document serves as both analysis and execution blueprint.
Executive Summary
Current State: daisyai.ai positions DaisyAI as a SaaS product ("AI-Powered Medical Case Review")
Target State: Position DaisyAI as a Forward Deploy Engineering (FDE) firm that builds production AI for healthcare, with the SaaS product serving as proof of capability
Why This Matters: DaisyAI is pivoting to a hybrid model (SaaS + FDE services). The website must reflect this to attract enterprise healthcare clients who need implementation partners, not just software vendors.
Part 1: Strategic Context
1.1 The FDE Thesis
The bottleneck for enterprise AI isn't models—it's implementation.
Key Statistics:
- 95% of AI projects fail to create business value (source: industry research)
- Only 1% of companies reach AI maturity
- 800%+ growth in demand for implementation engineers (Forward Deployed Engineers)
DaisyAI's Positioning:
- "Unlike consultants who deliver recommendations, we deliver working software"
- We embed with client teams, understand real workflows (not SOPs), and ship production code
- Our SaaS product proves we can build—it's running in production today
1.2 Target Market
Ideal Customer Profile:
- Regional health systems (3-20 hospitals)
- Specialty payers and TPAs
- UM/UR outsourcing companies
- Large physician groups with value-based contracts
- Revenue: $50M-$500M (big enough for budget, nimble enough to move fast)
Common Pain Points:
- Prior authorization backlogs
- Nurse burnout from manual review
- Denial management inefficiency
- Documentation bottlenecks
- Failed AI pilots with big vendors
1.3 Engagement Models
| Tier | Name | Price Range | Duration | Best For |
|---|---|---|---|---|
| 1 | Discovery Sprint | $15-25K | 2 weeks | "Is AI right for this workflow?" |
| 2 | Implementation Project | $75-200K | 2-4 months | "We know what we need, help us build it" |
| 3 | Embedded FDE | $20-30K/month | Ongoing | "Be our AI team" |
1.4 Competitive Differentiation
| Competitor Type | Their Approach | DaisyAI Difference |
|---|---|---|
| Big Consulting (McKinsey, Bain) | Strategy decks, 18-month timelines | We ship working code in weeks |
| Health IT Vendors (Epic, Cerner) | Monolithic platforms, long sales cycles | Lightweight, integrate with existing systems |
| AI Startups | Point solutions, no implementation support | Full-stack: strategy → code → change management |
| Staff Augmentation | Bodies without context | Domain expertise + engineering + product thinking |
Part 2: Research Findings
2.1 Palantir's Website (Inspiration)
What Works:
- "Get AI Into Operations" - action-oriented headline
- "Day 1 Value" messaging - emphasizes speed
- Third-party validation prominently displayed (Dresner, IDC, Forrester rankings)
- Educational content (AIP Bootcamps) builds trust
- Four-card layouts for feature presentation
- Clear product categories with dedicated pages
Apply to DaisyAI:
- Lead with action: "We Build AI That Actually Works"
- Emphasize speed: "Days, not years"
- Use stats as social proof (since no client logos yet)
- Education-forward approach via Substack content
2.2 SVPG's Website (Thought Leadership Model)
What Works:
- Team credentials front and center ("industry veterans with executive positions")
- Published books as credibility anchors
- Specific testimonials with concrete results
- Workshop offerings with dates (active engagement signals)
- Articles on major companies (Google, Spotify, Amazon) establish authority
Apply to DaisyAI:
- Highlight Wharton/Columbia/Princeton backgrounds
- Link to Substack articles as thought leadership
- Frame team as "left consulting to build"
2.3 BCG/McKinsey Style
What Works:
- Data-driven claims with research backing
- Heavy use of charts and statistics
- Pyramid Principle in content structure
- Professional, high-contrast design
Apply to DaisyAI:
- Lead sections with statistics
- Use the 95%/1%/800% stats prominently
- Structure content: insight → evidence → action
Part 3: Current Website Analysis
3.1 Tech Stack (Keep)
Framework: Next.js 15 + React 19
Styling: Tailwind CSS v4
Animation: Framer Motion
3D Effects: Three.js (Vortex background)
Fonts: Playfair Display (serif headings), Nunito Sans (body)
Deployment: Vercel (assumed)
3.2 Current Site Structure
/ # Homepage (SaaS-focused)
/about # Team page (basic bios)
3.3 Current Homepage Sections
- Hero - "AI-Powered Medical Case Review" (wrong message)
- ProblemStatement - Healthcare complexity (good, needs reframing)
- DaisySolution - Product features with mockup (good asset, wrong emphasis)
- Features - Product capabilities (move to /product)
- CallToAction - "Request Access" (should be "Book a Call")
3.4 Existing Components to Leverage
| Component | Location | Reuse Strategy |
|---|---|---|
Header.tsx | /src/components/layout/ | Modify: add nav + social links |
Footer.tsx | /src/components/layout/ | Modify: add social row |
Hero.tsx | /src/components/sections/ | Modify: new FDE messaging |
ScrollReveal.tsx | /src/components/ui/ | Keep: excellent animation |
Vortex.tsx | /src/components/backgrounds/ | Keep: distinctive visual |
RequestAccessModal.tsx | /src/components/ui/ | Rename to BookCallModal |
DaisySolution.tsx | /src/components/sections/ | Reference for new sections |
3.5 Design System (Keep + Extend)
Current Palette:
- Background:
#0a0a0a(near black) - Text:
whitewith opacity variants (/60,/70,/80) - Accent:
yellow-400(CTAs, highlights) - Borders:
white/10
Recommended Addition:
- Secondary accent: Consider
purple-500orteal-500for FDE-specific elements - Light section variant: For key conversion points (already exists in CallToAction)
Part 4: Target Site Architecture
4.1 New Page Structure
/ # Homepage (FDE-first narrative)
/about # Enhanced team + philosophy page
/services # NEW: Detailed engagement models
/product # NEW: SaaS platform showcase
/insights # NEW: Thought leadership hub (Substack integration)
4.2 New Navigation
Header (Desktop):
[Logo: Daisy AI] [Services] [Product] [Insights] [About] [Twitter] [LinkedIn] [Substack] [Book a Call →]
Header (Mobile):
[Logo] [☰ Menu]
4.3 Social Links
Platforms to Include:
- Twitter/X: @daisyai (or founder handle)
- LinkedIn: company page
- Substack: https://substack.com/@tstartz
Placement:
- Header: Small icons (16-20px), grouped, subtle
- Team section: Under each founder name
- Footer: Full row with labels
Part 5: Homepage Specification
5.1 Section Flow
1. Hero # Hook with FDE thesis
2. ProblemStats # NEW: 95% fail statistics
3. Approach # NEW: Embed → Build → Scale
4. EngagementModels # NEW: Three pricing tiers
5. ProofOfCapability # Modified: Product as evidence
6. TargetMarket # NEW: Who we work with
7. Team # Enhanced: Credibility focus
8. CallToAction # Modified: Book a call
5.2 Section Specifications
Section 1: Hero
Purpose: Hook visitors with FDE thesis in <5 seconds
Content:
Headline: "We Build AI That Actually Works in Healthcare"
Subhead: "95% of enterprise AI projects fail. We're the engineers who make them succeed."
CTA Primary: "Book a Discovery Call" → Calendly link
CTA Secondary: "See How We Work" → Scroll to Approach section
Design Notes:
- Keep Vortex background (distinctive)
- Large headline (text-5xl md:text-7xl)
- Stats in subhead create urgency
- Dual CTAs: conversion + engagement
Implementation:
- Modify existing
Hero.tsx - Keep scroll indicator animation
- Update headline/subhead content
- Change CTA button text and links
Section 2: ProblemStats (NEW)
Purpose: Establish credibility through market insight
Content:
Section Title: "The AI Implementation Crisis"
Stat 1: "95%" — "of AI projects fail to create business value"
Stat 2: "1%" — "of companies have reached AI maturity"
Stat 3: "800%" — "growth in demand for implementation engineers"
Body Copy:
"The bottleneck for enterprise AI isn't models—it's implementation.
Healthcare organizations have the budget and the pain, but lack the
talent to transform operations. Generic AI tools fail because they're
built against documentation, not how work actually gets done."
Design Notes:
- Dark background (continues from Hero)
- Large stat numbers with animation (count up on scroll)
- Stats in horizontal row on desktop, stack on mobile
- Body copy centered, max-width for readability
Implementation:
- New file:
/src/components/sections/ProblemStats.tsx - Use
ScrollRevealwrapper - Animate stats with Framer Motion
useInView - Reference
DaisySolution.tsxfor styling patterns
Section 3: Approach (NEW)
Purpose: Differentiate FDE from consultants and vendors
Content:
Section Title: "Unlike Consultants, We Build Working Software"
Column 1: "Embed & Discover"
- Shadow actual workflows
- Map the real process (not the SOP)
- Build trust with end users
- Identify highest-impact opportunities
Column 2: "Build & Deploy"
- Write production code on your infrastructure
- Integrate with existing EHRs and systems
- Ship iteratively, not in 18-month cycles
- Train your team as we build
Column 3: "Improve & Scale"
- Every engagement improves our platform
- Knowledge compounds across clients
- Your success feeds our product roadmap
- Graduate from projects to partnership
Design Notes:
- Three-column grid (stack on mobile)
- Icon for each column (suggest: Eye, Code, TrendingUp)
- Subtle hover effect on cards
- Consider numbered steps (1, 2, 3) for visual hierarchy
Implementation:
- New file:
/src/components/sections/Approach.tsx - Use existing card styling patterns
- Icons from Lucide React (already in project likely)
Section 4: EngagementModels (NEW)
Purpose: Make next steps concrete with pricing transparency
Content:
Section Title: "Three Ways to Work With Us"
Card 1: "Discovery Sprint"
Price: "$15-25K"
Duration: "2 weeks"
Includes:
- Process mapping & observation
- 10-15 stakeholder interviews
- Opportunity sizing & ROI model
- Working prototype demo
- Implementation roadmap
Best for: "Is AI right for this workflow?"
CTA: "Learn More" or expand
Card 2: "Implementation Project"
Price: "$75-200K"
Duration: "2-4 months"
Includes:
- Dedicated FDE team
- Production deployment
- Integration with existing systems
- Training & change management
- 30-day post-launch support
Best for: "We know what we need, help us build it"
CTA: "Learn More" or expand
Card 3: "Embedded FDE"
Price: "$20-30K/month"
Duration: "Ongoing"
Includes:
- Continuous engineering capacity
- Weekly syncs with leadership
- Quarterly business reviews
- Priority access to platform updates
- Expansion into new workflows
Best for: "Be our AI team"
CTA: "Learn More" or expand
Design Notes:
- Three cards, equal height
- Middle card could be "featured" (slight elevation or border)
- Pricing visible (transparency builds trust)
- "Best for" helps visitors self-select
- Consider expandable cards or link to /services
Implementation:
- New file:
/src/components/sections/EngagementModels.tsx - Card component with consistent structure
- Responsive: 3-col → 1-col on mobile
Section 5: ProofOfCapability (Modified)
Purpose: Show the SaaS product as evidence of building capability
Content:
Section Title: "We've Already Built It For Ourselves"
Intro: "This isn't vaporware—it's running in production."
Product Showcase:
- Simplified version of current DaisySolution mockup
- Key capabilities listed:
• 7 specialized AI agents for different review types
• Evidence extraction with source citations
• HIPAA-compliant, SOC2-ready infrastructure
• Integrates with existing clinical workflows
Closing: "We're not theorists. We're builders who ship."
CTA: "Explore the Platform →" links to /product
Design Notes:
- Reuse existing product mockup from DaisySolution
- Simplify feature list (detailed features go to /product)
- Frame as "proof" not primary offering
- Light background variant could work here for contrast
Implementation:
- Modify existing
DaisySolution.tsxor createProofOfCapability.tsx - Reduce content density
- Add link to /product page
- Rename component to reflect new purpose
Section 6: TargetMarket (NEW)
Purpose: Help visitors self-qualify, signal ICP
Content:
Section Title: "Built for Healthcare Operations Leaders"
Organization Types (with icons):
- Regional health systems (3-20 hospitals)
- Specialty payers and TPAs
- UM/UR outsourcing companies
- Large physician groups
Common Challenges:
✓ Prior authorization backlogs causing care delays
✓ Nurse burnout from repetitive manual review
✓ Denial rates impacting revenue cycle
✓ Documentation bottlenecks slowing decisions
Qualifier: "We work best with organizations doing $50M-$500M in revenue—
big enough to have budget, nimble enough to move fast."
Design Notes:
- Two-column layout: org types left, challenges right
- Icons for org types (Building, Shield, Clipboard, Users)
- Checkmarks for challenges (validation feeling)
- Qualifier text slightly smaller, helps filter
Implementation:
- New file:
/src/components/sections/TargetMarket.tsx - Icon grid + checklist layout
- Mobile: stack columns
Section 7: Team (Enhanced)
Purpose: Credibility through founder backgrounds
Content:
Section Title: "Wharton-Trained. Healthcare-Focused. Builder-Minded."
Founder 1: Thomas Startz
Role: CEO
Background: Columbia University, Wharton MBA
Experience: Healthcare operations and finance
Social: [LinkedIn] [Twitter]
Founder 2: Michael Yuan
Role: CTO
Background: Princeton University, Wharton MBA
Experience: Engineering and consulting (Bain, etc.)
Social: [LinkedIn] [Twitter]
Closing: "We've worked at Bain, McKinsey, and in healthcare operations.
We left to build something that actually ships."
CTA: "Read Our Thinking →" links to Substack
Design Notes:
- Side-by-side founder photos (existing assets)
- Social icons under each name
- Emphasis on "left to build" narrative
- Link to thought leadership
Implementation:
- Can extract from About page or create dedicated component
- New file if needed:
/src/components/sections/TeamHighlight.tsx - Reuse founder images from About page
Section 8: CallToAction (Modified)
Purpose: Convert visitors
Content:
Headline: "Let's Talk About Your AI Challenges"
Body: "Book a 30-minute call. No pitch, just problem-solving.
We'll tell you honestly if we can help—and if we can't, we'll point you
to someone who might."
CTA: [Book a Discovery Call] → Calendly
Alternative: "Prefer email? thomas@daisyai.ai"
Design Notes:
- Light background (invert from dark theme) - existing pattern
- Single prominent CTA button
- Email as fallback (some prefer async)
- Honest tone ("if we can't help...") builds trust
Implementation:
- Modify existing
CallToAction.tsx - Update copy
- Ensure Calendly link is correct
Part 6: Component Specifications
6.1 SocialLinks Component (NEW)
File: /src/components/ui/SocialLinks.tsx
Props:
interface SocialLinksProps {
platforms: ('twitter' | 'linkedin' | 'substack')[];
size?: 'sm' | 'md' | 'lg'; // 16px, 20px, 24px
variant?: 'header' | 'footer' | 'inline';
showLabels?: boolean;
className?: string;
}
URLs (to configure):
const SOCIAL_URLS = {
twitter: 'https://twitter.com/daisyai', // UPDATE with actual handle
linkedin: 'https://linkedin.com/company/daisyai', // UPDATE
substack: 'https://substack.com/@tstartz',
};
Design:
- Icons from Lucide or custom SVGs
- Header variant:
text-white/60 hover:text-white - Footer variant: larger, with optional labels
- Transition on hover
6.2 Header Modifications
File: /src/components/layout/Header.tsx
Changes Required:
- Add navigation links array
- Import and add SocialLinks component
- Update CTA text: "Request Access" → "Book a Call"
- Update CTA link to Calendly
New Navigation:
const NAV_LINKS = [
{ label: 'Services', href: '/services' },
{ label: 'Product', href: '/product' },
{ label: 'Insights', href: '/insights' },
{ label: 'About', href: '/about' },
];
Layout:
[Brand] [Nav Links] [Social Icons] [CTA Button]
6.3 Footer Modifications
File: /src/components/layout/Footer.tsx
Add:
- Full social links row with labels
- "Follow our thinking" or similar header
- Possibly newsletter signup (future)
Part 7: New Pages (Phase 2)
7.1 Services Page
Route: /services
File: /src/app/services/page.tsx
Content:
- Expanded engagement model details
- FAQ for each tier
- Process timeline visualizations
- "What to expect" sections
7.2 Product Page
Route: /product
File: /src/app/product/page.tsx
Content:
- Migrated from current homepage
- Full feature showcase
- Agent descriptions (7 specialized agents)
- Screenshots/mockups
- Integration details
- Pricing (if separate from services)
7.3 Insights Page
Route: /insights
File: /src/app/insights/page.tsx
Content:
- Substack embed or curated article list
- Categories: FDE, Healthcare AI, Implementation
- Links to external publications
Part 8: Implementation Checklist
Phase 1: MVP Homepage (Priority)
Files to Modify:
-
/src/app/page.tsx- Restructure section composition -
/src/components/sections/Hero.tsx- New FDE messaging -
/src/components/layout/Header.tsx- Nav links + social + CTA -
/src/components/layout/Footer.tsx- Social links row -
/src/components/sections/CallToAction.tsx- Updated copy
Files to Create:
-
/src/components/ui/SocialLinks.tsx- Reusable social icons -
/src/components/sections/ProblemStats.tsx- Stats section -
/src/components/sections/Approach.tsx- Three pillars -
/src/components/sections/EngagementModels.tsx- Pricing cards -
/src/components/sections/TargetMarket.tsx- ICP section -
/src/components/sections/TeamHighlight.tsx- Founder section
Content Needed:
- Finalize headline/subhead copy
- Verify stat sources (95%, 1%, 800%)
- Confirm pricing ranges are accurate
- Get social media URLs (Twitter, LinkedIn)
- Confirm Calendly link
Phase 2: Supporting Pages
-
/src/app/services/page.tsx -
/src/app/product/page.tsx -
/src/app/insights/page.tsx - Update navigation links
Phase 3: Polish
- Mobile responsive testing
- Accessibility audit
- SEO meta tags
- Analytics events
- Performance optimization
Part 9: Design Decisions
Keep Dark Theme + Vortex
Rationale:
- Differentiation - Most healthcare sites are sterile blue/white
- Premium positioning - Dark = high-end, sophisticated
- Technical credibility - "We're engineers, not marketers"
- Memorable - Vortex creates distinctive first impression
Show Pricing Publicly
Rationale:
- Transparency builds trust - No "contact us for pricing" games
- Qualifies leads - Filters out wrong-fit inquiries
- Demonstrates confidence - We know our value
- Aligns with FDE ethos - Direct, no BS
Direct Calendly (No Form)
Rationale:
- Reduces friction - One click to book
- Shows availability - Signals responsiveness
- Self-service - Respects visitor time
- Can add form later - Start simple, optimize based on data
Part 10: Open Items
Needs Owner Input
-
Social Media URLs
- Twitter handle: _______________
- LinkedIn company page: _______________
- Substack confirmed: https://substack.com/@tstartz
-
Calendly Link
- Current: https://calendar.app.google/aVV7SrwdnLTsvxyL6
- Confirm this is correct for "Book a Call"
-
Pricing Confirmation
- Discovery Sprint: $15-25K (confirm)
- Implementation: $75-200K (confirm)
- Embedded: $20-30K/month (confirm)
-
Stat Sources
- 95% AI projects fail - source?
- 1% AI maturity - source?
- 800% FDE growth - source?
-
Founder Details
- Thomas: Additional credentials to highlight?
- Michael: Specific companies to name (Bain confirmed)?
Appendix A: File Structure Reference
daisyai-www-v2/
├── src/
│ ├── app/
│ │ ├── page.tsx # Homepage (MODIFY)
│ │ ├── about/
│ │ │ └── page.tsx # About page (MODIFY)
│ │ ├── services/
│ │ │ └── page.tsx # NEW
│ │ ├── product/
│ │ │ └── page.tsx # NEW
│ │ └── insights/
│ │ └── page.tsx # NEW
│ │
│ ├── components/
│ │ ├── layout/
│ │ │ ├── Header.tsx # MODIFY
│ │ │ └── Footer.tsx # MODIFY
│ │ │
│ │ ├── sections/
│ │ │ ├── Hero.tsx # MODIFY
│ │ │ ├── ProblemStats.tsx # NEW
│ │ │ ├── Approach.tsx # NEW
│ │ │ ├── EngagementModels.tsx # NEW
│ │ │ ├── ProofOfCapability.tsx # NEW (or modify DaisySolution)
│ │ │ ├── TargetMarket.tsx # NEW
│ │ │ ├── TeamHighlight.tsx # NEW
│ │ │ └── CallToAction.tsx # MODIFY
│ │ │
│ │ ├── ui/
│ │ │ ├── SocialLinks.tsx # NEW
│ │ │ └── ... (existing)
│ │ │
│ │ └── backgrounds/
│ │ └── ... (keep as-is)
│ │
│ └── lib/
│ └── ... (keep as-is)
│
└── public/
└── images/
└── ... (add any new assets)
Appendix B: Prompt for Execution
When starting work on this refactor, use this prompt:
I need to refactor the DaisyAI website (daisyai-www) from SaaS-focused to FDE-first positioning.
Read the specification at: [path to this file or paste content]
Key context:
- This is a Next.js 15 + React 19 + Tailwind v4 site
- Keep the dark theme and Vortex background
- The goal is FDE-first: lead with consulting services, SaaS product is proof of capability
- No existing case studies - use product as evidence of building capability
- Need social links (Twitter, LinkedIn, Substack) in header and footer
Start with Phase 1 (MVP Homepage):
1. Create the SocialLinks component first
2. Modify Header with nav links and social icons
3. Create new sections: ProblemStats, Approach, EngagementModels, TargetMarket, TeamHighlight
4. Modify Hero with new FDE messaging
5. Update page.tsx to compose the new section order
6. Modify CallToAction with updated copy
Maintain the existing design system and animation patterns. Reference existing components like DaisySolution.tsx and ScrollReveal.tsx for styling consistency.
Document created: January 8, 2025 Version: 1.0 Author: Claude Code (OS repo session)