Video summary

Why AI-Generated Websites Don’t Convert

Main summary

Key takeaways

Product Review

Product / Website Reviewed

An AI-generated landing page for a fictional AI fitness coaching app called Pulse, built in Codeex using a GPT-6 Astro model.

The video explains why AI-generated websites that look good can still fail to convert, and demonstrates conversion-focused redesign tweaks.

Key Features Mentioned (as originally built)

  • Polished, consistent visual design and branding
  • Clean hero with product imagery, UI showcase cards, and scroll effects
  • Dynamic carousel section
  • Testimonials section
  • Pricing card with a signup prompt
  • FAQ with dynamic accordions
  • Multiple signup CTAs across the page
  • Several “Get started / join pulse / start training”-style buttons that lead to the same signup flow

Main Issues Hurting Conversion (before redesign)

  • Vague hero messaging
    • Example: “Find your rhythm”
    • Didn’t clearly explain:
      • what the app does
      • what the user gains
  • Hero visual not conversion-focused
    • Right-side graphic was too generic
    • Product value wasn’t clearly demonstrated (e.g., small progress indicator)
  • Too many competing CTAs / unclear CTA wording
    • Multiple “primary” actions (e.g., Get started, Join Pulse) plus an extra Explore button competing for attention
  • Free trial form had too much friction
    • Multi-field form (e.g., goal, experience level, workouts per week) likely caused drop-off
  • Consent & tracking overlooked in the original demo setup
  • Outcomes/benefits lacked clarity
    • The page showed features more than concrete results
  • CTA visibility problem
    • After the hero section, there were long stretches without an obvious free trial button
  • Weak trust signals
    • Testimonials felt generic/template-like
    • Missing stronger, more credible proof (e.g., app-store-style metrics)

Conversion-Focused Changes Made (redesign)

  1. Improved hero clarity

    • Updated headline/subheadline to be specific and value-driven
    • Tied adaptive workout plans to schedule/equipment/goals
    • Redesigned the right-side graphic with more product-specific UI examples
      • e.g., “Wednesday adjusted lighter mobility session” chip

  2. Reduced CTA competition + clarified CTA wording

    • Removed the secondary competing button
    • Standardized primary CTAs to “Start my free trial”
    • Added supporting context next to the primary CTA
      • e.g., “14 days free, cancel anytime”

  3. Lowered signup friction

    • Replaced multi-field signup with a single email input
    • After email submission, the user sees:
      • the official App Store download button
      • a QR code option for desktop users

  4. Added consent + measurement compliance using Cookiebot

    • Used Cookiebot (Usercentrics) to scan cookies/trackers and provide a consent banner
    • Supports analytics/ads behavior via:
      • Google Consent Mode
      • Microsoft Consent Mode
    • The video emphasizes this corrects what the original AI-generated site overlooked

  5. Replaced generic section with outcome proof

    • Swapped a generic background section for a concrete stat:
      • “Members average 3.2 sessions a week after month one.”
    • Reused the bright blue background style from the CTA area so the stat stands out

  6. Kept the CTA accessible at all times

    • Made the navbar sticky so the free trial CTA remains visible
    • Mobile behavior:
      • hamburger menu
      • primary CTA on the right

  7. Upgraded trust signals from generic testimonials to app metrics

    • Replaced template testimonials with:
      • App Store rating: 4.8 / 5
      • Over 10,000 real downloads

Numerical Ratings / Scores Explicitly Stated

  • 4.8 out of 5 stars (app rating)
  • Over 10,000 real downloads
  • 3.2 sessions a week after month one (outcome stat)

Pros (as demonstrated)

  • Much clearer value proposition and product explanation (especially in the hero)
  • Less-friction signup flow (email-only + official App Store button + QR)
  • Stronger trust using concrete, app-store-like metrics instead of generic testimonials
  • Better measurement readiness through consent handling (Cookiebot + consent mode)
  • Improved conversion UX by keeping the CTA visible (sticky navbar)

Cons / Limitations Mentioned (implicit)

  • The original AI version’s biggest weaknesses weren’t aesthetic—they were business/UX/measurement issues.
  • The redesigned CTA still requires users to start a free trial (the form is simpler, but still assumes intent and willingness).

Comparisons Made

The video contrasts:

  • AI-demo-looking polish vs conversion-focused design
  • Feature-focused messaging vs outcome-focused messaging
  • Template testimonials vs credible app store rating + download numbers
  • An “AI-generated website” that overlooks tracking/consent vs a setup that includes Cookiebot compliance

Overall User Experience Impact (from the demo narrative)

  • First impression improves because hero messaging immediately explains:
    • “what it is”
    • “why it matters”
  • Signup becomes faster and less intimidating (fewer fields)
  • Users can always find the CTA (sticky navbar)
  • Trust increases through more concrete, verifiable-like stats

Unique Points (consolidated)

  • An AI-generated site may look convincing yet still underperform on conversion
  • Conversion depends on:
    • CTA visibility/clarity
    • low-friction path to action
    • clear value for the right audience
    • trust signals
    • measurement via analytics
  • Original hero was more decorative than conversion-focused
  • Multiple/competing CTAs reduce clarity
  • CTA copy should be explicit (e.g., “Start my free trial” vs “Get started”)
  • Signup form was too long (multi-field → higher drop-off risk)
  • Consent/tracking requires proper setup (Cookiebot + consent banner + Google/Microsoft consent mode)
  • Add an “outcomes” section with concrete results (e.g., 3.2 sessions/week after month one)
  • Make CTA constantly accessible via sticky navbar
  • Replace generic testimonials with stronger proof:
    • 4.8/5 stars
    • 10,000+ downloads
  • Design changes should be subtle but purposeful (nearly the same layout after redesign)

Speaker Views / Aspects (if multiple)

  • Single main speaker: focuses on a conversion audit framework + practical redesign steps
  • Sponsored element: Cookiebot integration is presented as the fix for tracking/consent

Verdict / Recommendation

AI-generated landing pages shouldn’t stop at visual polish. The video recommends auditing conversion blockers such as:

  • hero clarity
  • CTA strategy
  • signup friction
  • consent/tracking setup
  • outcome messaging
  • credible trust signals

For Pulse-like landing pages, the redesign approach—especially explicit CTA wording, email-only signup, consent compliance, outcome stats, and app-store rating/download proof—is presented as the recommended direction and should materially improve conversion.

Original video