Median readiness score
69
Current sample median where score data is present
2026 v1 report
The State of AI-Built Websites 2026.
PageLens AI turns launch scans into a public research view: what AI-built and fast-shipped websites get right, what they miss, and which fixes make them safer to show the internet.
Current findings
The report PageLens AI can own is already live.
Ahrefs and Semrush can tell teams where traffic might come from. PageLens AI can tell them whether the site is credible, accessible, trackable, secure and understandable enough to receive that traffic.
Snapshot generated 23 Jul 2026. The sample is directional and limited to PageLens AI completed scans, not the whole web.
Score 90+
8%
Scans classed as excellent in the benchmark sample
Score below 70
34%
Scans needing meaningful launch-readiness work
Failed scan rate
0%
Recorded failed runs versus completed or failed runs
What the report tracks
The sections map to launch blockers, not vanity metrics.
Each section connects a visible website failure to the owner decision it affects: trust, traffic, conversion, safety or client handoff.
Most common launch blockers across AI-built sites
Average launch-readiness score by builder and site type
Trust-signal failures, missing proof and unclear ownership
AI Search readiness gaps: structure, clarity and citeability
Tracking, analytics, consent and cookie mistakes
Before-and-after improvements after PageLens AI rescans
Builder segments
The benchmark follows the way people actually build now.
The report is organised around launch workflows, not old software categories. That makes the findings useful to founders, agencies, AI-builder communities and the platforms themselves.
AI app builders
Lovable, Bolt, Cursor, Codex, Claude Code, v0, Replit and other fast-build workflows where the product looks finished before the public surface is reviewed.
No-code and visual builders
Webflow, Framer, Carrd, Notion sites, Wix and Squarespace launches where polish, metadata, mobile UX and accessibility often diverge.
Commerce launches
Shopify, WooCommerce and BigCommerce stores where trust, tracking, product-page clarity and checkout confidence decide whether traffic converts.
Measured signals
The headline metrics come straight from launch blockers.
Each signal maps to a visible failure a buyer, crawler, AI answer engine, community reviewer or first customer could notice.
Missing or weak meta descriptions
Broken Open Graph previews
Poor mobile tap targets
Missing structured data
Weak trust proof and unclear audience
Security header gaps
Consent, cookie and analytics mismatches
Unclear AI-answerability and citation cues
Research engine
The weekly teardown flywheel feeds the annual report.
Teardowns create useful media every week. The report turns those scans and rescans into category authority over time.
- 1
Sample recent completed public-site scans from the PageLens AI dataset.
- 2
Aggregate issue categories, severity and anonymised fix patterns.
- 3
Publish permissioned examples through Weekly Teardown.
- 4
Track rescans so the report can show which fixes move readiness scores.
Want your site in the evidence layer?
Run a scan, fix the report, verify your domain and choose whether your public trust page can be listed. The research layer stays aggregate unless a site owner opts into public proof.