The starting point: the specialist's own site

Surge45 is an AI search intelligence and advisory firm for SaaS. It already did much of what it recommends: a robots.txt that names and allows the major AI bots, and real llms.txt files, the only ones among the five brands in these stories. The gaps the audit found were in speed and structure, the details that are easy to miss on your own site.

What the AI Discoverability audit found

WriteWorks audited a sample of surge45.com. The approximate readiness score was already excellent.

  • Every page was rendered on request. Every audited HTML response was uncached (private, no-cache, no-store). That made Surge45 the slowest of the five sites, with only 75% of pages answering in under 0.8 seconds at their worst timing.
  • Heading structure broke the answer-first pattern. Only 70% of audited pages passed the answer-first check. The main cause was skipped heading levels (such as an h2 followed by an h4): only 32% of articles and FAQs had a clean heading structure.
  • Some FAQ blocks didn't parse. Unescaped quotes broke FAQPage markup on some guides, so only 87% of FAQ blocks parsed.
  • The sitemap listed URLs that weren't final. Only 84% of the audited sitemap URLs were live, indexable pages; the rest redirected to new paths or were noindex thank-you pages.

What changed after the fixes

  • Resolved: caching. Every audited response now allows caching (0% → 100%), and 87% were served straight from cache, up from 0%. The median server response is 2.5 times faster.
  • Improved: slow pages. Pages answering in under 0.8 seconds rose from 75% to 86%, and none took longer than 1.8 seconds. Nearly all the slow pages left are insight articles not yet served from cache, so they are next.
  • Resolved: headings. Articles and FAQs with a clean heading structure rose from 32% to 100%, and pages passing the answer-first check from 70% to 100%.
  • Resolved: FAQ markup. FAQ blocks that parse rose from 87% to 100%.
  • Resolved: the sitemap. Audited sitemap URLs that are live, indexable pages rose from 84% to 100%.

How these numbers were measured

Before = the first WriteWorks AI Discoverability audit, after = the re-audit once the fixes were deployed, on the same pages with the same rules. Percentages are shares of the audited pages, not of every URL on the site. A fix is marked Resolved when the re-audit shows it fixed, Improved when it is partly fixed, and Open when nothing changed. Speed comparisons come from one vantage point and are indicative. AI platform coverage is the share of the AI platforms measured that cited the site, at the baseline and six months on: a timeline, not a claim about cause. One rule result is left out of this story: the audit flagged homepage Organization markup as missing, but the homepage carries equivalent business markup with sameAs links, so that flag is a false positive in the rule rather than a site issue.