A clinical workflow SaaS ranked well on Google and was almost never mentioned when buyers asked ChatGPT, Claude or Perplexity which vendors to consider. The gap was entirely fixable and almost entirely self-inflicted.
The company sold clinical workflow software to hospital operations and clinical informatics teams. Traditional search performance was good — the site ranked on the first page for most of its priority commercial terms and organic traffic was healthy.
The problem surfaced in sales calls. Prospects were arriving with a shortlist of three or four vendors, assembled by asking an AI assistant which clinical workflow platforms to consider. This company was almost never on that list, despite ranking above several vendors who were.
The baseline audit made the causes clear. First, a security plugin at CDN level was blocking GPTBot, ClaudeBot and PerplexityBot — nobody had configured this deliberately and nobody knew. Second, the company described itself inconsistently across fourteen sources: the website, LinkedIn, three directories, two review platforms, a partner page and several press releases each gave a different founding year, employee count, product category or headquarters location.
Third, the content was written as marketing copy rather than as reference material. Pages opened with three paragraphs of positioning before stating what the product did, which is exactly the structure that makes content hard to extract and cite.
Figures measured over the engagement period and reconciled against the client CRM.
| Metric | Baseline (month 1) | Month five |
|---|---|---|
| Mention rate across 60 prompts | 8% | 61% |
| ChatGPT mention rate | 7% | 64% |
| Claude mention rate | 5% | 57% |
| Perplexity citation rate | 13% | 71% |
| Factual errors in AI descriptions | Frequent — category and size wrong | Rare |
| External sources with contradictory facts | 14 | 0 |
Mention rate is the percentage of a fixed 60-prompt set in which the company was named or cited, re-run monthly under consistent conditions. Because assistant outputs are non-deterministic, these figures are directional rather than precise.
Entity consolidation, crawler access, cornerstone content and monthly prompt tracking.
Schema deployment and answer-format restructuring for citation extraction.
The organic foundation the assistants were retrieving from.
Third-party corroboration through directories, reviews and publication presence.
Supplied the factual specificity and category definition the entity work required.
Landing experience adjusted for visitors arriving further along in research.
With a fixed prompt set. We built sixty questions this company's buyers actually ask, ran them monthly across ChatGPT, Claude, Gemini and Perplexity under consistent conditions, and recorded whether the company was named, which competitors were named, and whether the description was accurate. Single screenshots are not evidence because outputs are non-deterministic.
Three things in order of impact. AI crawlers were blocked at CDN level by a security plugin nobody had configured deliberately. The company described itself inconsistently across fourteen external sources. And content was written as marketing copy rather than extractable reference material.
Roughly a third, and it took one day. This is why crawler access is the first thing we check in every audit — it is free to fix, total in its effect when broken, and a surprising number of sites are blocking AI crawlers without knowing it.
No, and this engagement is the clearest example we have. The company outranked several competitors who were routinely named in assistant answers while it was not. Ranking depends on relevance and authority; being recommended depends additionally on retrievability, entity clarity and third-party corroboration.
Neither, and we say so in every scoping conversation. Assistant outputs are non-deterministic and vary by user, region and model version. This engagement had an unusually large and cheap fix available in the blocked crawlers. Companies without that problem see slower, smaller improvements from the same work.
Figures on this page are reconciled against the client CRM at the end of the engagement window. Client identity withheld at their request.
The free AI visibility audit builds a prompt set for your category, runs it across four assistants, and tells you your baseline mention rate and what is causing it.
+91 93266 17504 · India · USA · UAE
Data-Driven B2B Growth Marketing. Elevate Lead Generation, ABM & Demand Generation
