Case study: #2 in ChatGPT on its niche… and invisible to its buyers

We audited the AI visibility of a B2B travel-tech SaaS (with their permission — the brand is anonymized). The result fits in one sentence: the AI engines love this company when you speak to them in its own words, and completely ignore it when you speak in its buyers' words. Here are the numbers, the explanation, and the action plan.

The context

Our client is a young SaaS platform that automates a critical operational problem for enterprise customers — a niche B2B market dominated by well-established incumbents. Its buyers are operations directors: busy people who increasingly ask their buying questions to ChatGPT or Perplexity rather than Google.

The founder's question was simple: "When a buyer asks an AI which solutions exist in my market, do we show up?"

The method

We queried four engines (ChatGPT, Perplexity, Gemini, Claude) with two sets of questions:

• The niche vocabulary — the exact terms the company uses to describe what it does;
• The buyers' vocabulary — the market's generic terms: "best [category] software", "top tools for [problem]", comparison questions between established solutions.

In total: 20 answers analyzed on buyer queries, plus a niche-specific prompt set, plus a technical audit of the site.

Finding #1: on its exact niche, the company already exists — convincingly

When the question matches the company's precise positioning, it comes out #2 in AI visibility (35.7%), tied with the sector's historic leader — a publicly traded giant. For a company just a few years old, that's remarkable. Its site is even already cited as a source by the engines, and the sentiment expressed by the AIs is 69% positive.

Finding #2: on the buyers' vocabulary, zero

Then we asked the questions the way a buyer actually asks them — "best [category] software", "top tools for [problem] in 2026", "comparison of [leader A] vs [leader B]". Result: 0 mentions out of 20 answers. None of the four engines cites the company, on any question. The answers are monopolized by four or five incumbents — always the same ones.

Why the gap?

Generative engines don't "rank" sites the way Google does: they cite the sources that answer the question most clearly, in the words of the question. The company describes its product in its own language — precise, differentiated, but nobody types it. Its buyers use the sector's standard jargon. Between the two: a vacuum, which competitors fill.

The technical audit confirmed the weak foundations: no structured data (JSON-LD), no llms.txt file, an inaccessible sitemap, a single ranked keyword on Google and zero presence in Google AI Overviews. The engines have very little official material to read — so they read everyone else.

The action plan (6 priorities)

1. Technical quick wins (one day): JSON-LD structured data, an llms.txt file, an accessible sitemap.
2. Occupy the buyer vocabulary: pages and articles that answer questions the way buyers phrase them — not the way the company names its product.
3. Publish the comparison pages buyers already ask the AI engines about ("solution A vs solution B") — and become the source that answers them.
4. Earn citations in the media the AI engines read: the sector's trade press, LinkedIn, reference directories.
5. Neutralize the reservations the AI engines repeat (every expressed doubt = a proof page to publish: case studies, numbers, documentation).
6. Prioritize the weakest engine — here Perplexity (8%), which feeds on fresh web citations: it's the biggest gap, therefore the fastest gain.

The lesson, for any B2B company

Your next customer may never see your website: their AI already has an opinion about your market. It recommends three or four names, confidently, with no page two and no ads. And it doesn't pick the "best": it picks those who published the clearest content, in the words buyers use.

The good news from this case study: it's winnable, fast. A young company can sit level with a publicly traded giant — our client proves it on its niche. What's missing isn't credibility: it's the bridge between its vocabulary and its market's.