A Third of Your AI Visibility Is Happening in Territory Your Team Has No Instrument to Measure.

Your content team is tracking the wrong keywords.

AirOps just published one of the most rigorous studies I've seen on how ChatGPT actually decides what to cite — 548,534 pages analysed, 43,233 queries, across 8 intent categories.

The findings should change how boards think about AI-era visibility. Whether you're running a private bank, a luxury maison, or any brand where trust and authority drive revenue.

The number that stopped me: 32.9% of everything ChatGPT cited appeared only in follow-up searches that ChatGPT generated on its own — not in the original query someone typed. And nearly 95% of those follow-up searches showed zero monthly volume in standard SEO tools.

Let that sink in. A third of your AI visibility is happening in a space your marketing team literally cannot see.

How ChatGPT Actually Searches (And Why It Matters for Your Brand)

When someone asks ChatGPT "Which private banks offer the best wealth management for UHNW clients?" or "What luxury brands are leading in sustainability?", it doesn't search that phrase once and stop.

It expands. Generating multiple follow-up queries — what AirOps calls "fan-out" — to fill gaps, verify assumptions, and find alternative angles. This happens on 89.6% of all searches. The queries it generates are specific, long-tail, and invisible to any keyword tool built before 2024.

This isn't an edge case. It's how the product works.

If your content strategy is built entirely on monthly search volume — and most luxury and financial services brands' strategies still are — you are optimising for a world that no longer describes how AI-mediated discovery works.

For a bank CEO: your team is measuring 67% of the map. The other 33% is where ChatGPT is forming its opinion about whether to recommend you — and you have no data on it.

For a luxury CEO: your brand narrative is being shaped by queries your team has never seen, pulling from sources your team may never have audited.

What Google Rankings Still Get Right

The study also confirms something important: existing SEO investment still matters.

Pages ranked #1 on Google were cited by ChatGPT 43.2% of the time — 3.5x more often than pages outside the top 20.

But that's a floor, not a ceiling. Google rank is necessary. It's no longer sufficient.

And here's the finding most boards will miss: sites with domain authority below 80 received 74% of ChatGPT citations — compared to just 25.4% from the highest-authority domains. The platforms your team may have written off as too small to compete with are capturing more AI citations than your biggest competitors.

The reality is that the playing field for AI-era visibility is more level than the SEO era suggested. Smaller, more specific content wins more often than brand weight alone. A well-written article on a niche finance publication about your wealth management approach can outperform your own corporate website in ChatGPT's citations.

For luxury brands facing reputation challenges — lab-grown diamond debates, sustainability scrutiny, pricing backlash — this means critical third-party content on mid-authority sites is shaping AI perception far more than your own polished communications.

The Easiest Lever Hiding in Your Existing Content

Here's where this gets immediately actionable.

Pages where the title overlapped with the query by more than 50% were cited 2.2x more often than lower-alignment pages. No new content needed. No technical investment. Title optimisation on existing pages — particularly those sitting at positions 10-20 on Google, where the citation advantage strengthens with refresh.

This is a lever most content teams haven't been told to pull yet.

And it compounds: only 15% of pages ChatGPT retrieved were ultimately cited — meaning 6 out of 7 pages that passed initial discovery still failed to earn a citation. Being found is necessary but far from sufficient. The gap between retrieval and citation is where most organisations have no visibility or strategy.

Citation rates also varied dramatically by query intent:

→ Product-discovery queries: 18.3% citation rate → How-to queries: 16.9% citation rate → Comparison queries: 13.1% citation rate → Validation queries: 11.3% citation rate

Content optimised for bottom-of-funnel commercial intent — the kind most luxury and financial services brands invest in — is the hardest to get cited by AI. Content that explains, teaches, and demonstrates expertise earns citations nearly twice as often.

For banks: your product pages are almost invisible to ChatGPT. Your thought leadership on market trends and wealth strategy? That's what gets cited.

For luxury brands: your collection lookbooks don't earn AI citations. Your brand heritage content, sustainability deep-dives, and craftsmanship stories do.

Three Questions for Any Board — No Technical Expertise Required

If this research tells us anything, it's that AI visibility is now a governance-level issue, not a marketing team deliverable. Here are three questions any CEO can ask this quarter:

→ "What are we doing to track AI citations, not just organic traffic?"

Most analytics dashboards were not built to measure AI-referred visibility. If your team can't answer this, you have a measurement gap before you have a strategy gap. Start there.

→ "Are our titles written for search queries or for brand language?"

Title-query alignment is a content discipline, not a technical one. The 2.2x lift comes from specificity — titles that mirror how people actually ask questions, not how your brand describes itself internally.

→ "Which pages sit at positions 10–20 on Google and haven't been refreshed in the past year?"

This is where citation investment compounds fastest. Not net-new content — better-maintained existing content with titles that reflect current search behaviour.

The Bottom-line:

AI is increasingly the first touchpoint between your organisation and a client, partner, or investor running research. Gartner projects organic search traffic drops 50% by 2028. The shift is structural and accelerating.

The companies that treat this as a measurement and content discipline now — before it becomes standard practice — will compound visibility while others are still deciding whether to act.

The data is public. The methodology is clear. The question is whether your organisation is measuring what actually matters in 2026 — or still optimising for a map that's missing a third of the territory.

---

Source:

• AirOps AI Visibility Report: The Influence of Retrieval, Fan-out, and Google SERPs on ChatGPT Citations — 548,534 pages, 43,233 queries, 8 intent categories
• Gartner: Organic Search Traffic to Drop 50% by 2028