MethodologyAI readiness
How we score AI readiness
The AI readiness scan answers one question: can an AI shopping agent actually read a store, and what does it find when it does. Every score is computed by the same automated pipeline, no manual grading, including the one on this page.
Three phases, in order
Each scan runs the same three phases against the target site. Nothing here is inferred from a company's reputation or category; it's read directly off the live pages.
Crawler access
We fetch robots.txt and check fourteen tracked AI bot user-agents across three categories: training crawlers, AI search crawlers, and real-time shopping agents, against the site's allow/disallow rules.
Discoverability
We fetch the XML sitemaps, count URLs and lastmod coverage, and check for an llms.txt file at the standard location.
Page sampling
We fetch up to ten product pages twice: once as raw HTML, once through a real browser. We extract structured data from each, score completeness across ten product fields, and compare what an AI crawler sees against what a person with JavaScript sees.
Five signals, weighted, added up
The three phases above feed five weighted signals. Nothing is averaged unevenly or adjusted by hand; the weights are fixed in code and sum to 100%.
Laup.ai's own scan, 2 July 2026:
What the scan actually found on laup.ai
A few of the concrete signals behind the score above. The full interactive report, including per-page detail, is linked below.
Crawler access: fully open
All fourteen tracked AI bots are allowed by robots.txt, including GPTBot, ClaudeBot, Google-Extended, and PerplexityBot.
Sitemap: complete
All nine sitemap URLs carry a lastmod date, and an llms.txt file was found at the standard location.
Content visibility: SSR gaps
The pricing page rendered 437 characters of text in raw HTML versus 963 once JavaScript ran, the exact SSR/CSR gap this scanner is built to catch.
Product schema: zero, honestly
Laup.ai is a marketing site, not a product catalog, so no page carries product schema. The score counts that as a gap rather than skipping it.
Common questions
What does the AI readiness score measure?
It measures whether an AI shopping agent can actually read and understand a site's pages, weighted across five signals: content visibility (40%), product schema completeness (30%), crawler access via robots.txt (15%), sitemap discoverability (10%), and homepage schema (5%). The weights sum to 100% and the same weighted formula runs on every scan.
Why would a real site score less than 100?
Missing data counts as zero on purpose. Laup's own scan of laup.ai scores zero on product schema completeness, because laup.ai is a marketing site, not a product catalog, so there is no product schema for the scanner to find. That is exactly the kind of gap the score is built to surface: a signal that is quietly absent, not a score anyone had to manually assign.
What counts as blocking an AI crawler?
A robots.txt rule that denies a tracked AI bot access to the site root, across three categories: training crawlers (GPTBot, ClaudeBot, Google-Extended, and others), AI search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot), and real-time shopping agents (ChatGPT-User, Claude-User, Google-Agent, Perplexity-User). Only search and real-time access count toward the score; blocking training crawlers is treated as a legitimate choice, not a penalty.
How often is a scan re-run?
Anyone can re-run a scan at any time by submitting a URL. The sample report on this page reflects a specific scan of laup.ai and is not live-updated; visit the linked report for the full detail behind these numbers.
Next chapter
How we measure AI Visibility
This chapter covers whether AI can read your site. The next one covers whether AI actually recommends it.
See your own score
Run the same scan on your site. Takes about thirty seconds, no account needed to see the result.