Case study · 23 July 2026
We ran Scout on our own studio. It hurt. Then it worked.
instlytechnologies.com - the site of the company that builds Scout - audited with the same scanner our customers use. Every number below is dated, real, and re-checkable.
What Scout found
Three findings we'd have sworn weren't there.
HIGHZero structured data - on an AI infrastructure company's website
Not one JSON-LD block. Machines reading the site had no Organization entity, no WebSite node, nothing to anchor a citation to. The company that audits AI readability was itself invisible to the machines. Fix: a proper Organization / WebSite / WebPage schema graph, deployed the same afternoon.
HIGHFonts blocking first paint - ~1.9 seconds of invisible page
The Google Fonts stylesheet loaded render-blocking, and a 56 KB Supabase bundle - 82% unused on the public site - was competing with CSS for bandwidth before anything painted. Fix: fonts injected asynchronously with a no-JS fallback, the unused bundle deferred out of the critical path. Lighthouse mobile went from a consistent 69 to a ~90 median (Google PageSpeed API, four runs).
HIGHConversion paths only machines couldn't take
Every call-to-action was a JavaScript-only button - "Start a conversation" went nowhere a crawler or AI agent could follow. Fix: CTAs became real links (with a mailto fallback) that still open the same contact panel for humans. Agents now see a genuine conversion destination.
And one finding where Scout itself was wrong
Scout flagged our logo images as "unlabeled". They weren't - they used the correct accessibility pattern for decorative images (empty alt, aria-hidden, labeled wrapper). We fixed the scanner, not the site: Scout now recognises WCAG-correct decorative markup everywhere. When our own tool is wrong, we say so and patch the engine - your audit gets the benefit.
How the day went
One day, in order.
Morning: surface x-ray showed SCHEMA 0 blocks and two "unlabeled" images on instlytechnologies.com. Full extension scan: 68/100 - the findings above, ranked by impact and effort.
Midday: schema graph written and deployed; fonts made asynchronous; the unused bundle deferred; CTAs converted to crawlable links. Each fix verified live before moving on.
Afternoon: re-scan. AI Visibility 68 → 78. Google PageSpeed 69 → ~90 median. Surface x-ray: all seven checks green. The remaining gap to the 90s is content depth - named proof and answer pages - which is exactly what the report says, because that's what the machines say.
Verify any of it: run the free scan on instlytechnologies.com yourself, or check Google PageSpeed Insights directly. The audit methodology is the same one in the sample report.
Your site has findings like these. Every site does.
Free surface scan in seconds. The full audit ranks every fix by impact - like it did for us.