Ask a business owner what AI thinks of their business and you’ll usually get one of two answers: a blank stare, or a number. Some tool gave them a score — domain authority, an AI visibility grade, a semicircular gauge sitting at 62 — and now they’re either reassured or panicking about a metric nobody can define. Here’s the problem with that number: there isn’t one. No AI search system publishes a universal “authority score” for your business. What exists instead is evidence — accumulated, messy, spread across the web — and the systems deciding, answer by answer, whether your business is worth retrieving, citing, or naming. That’s a very different problem from raising a score, and honestly, it’s better news. Evidence can be built. A proprietary gauge just has to be fed.

What Google actually says

Start with the only source that gets to define its own system. Google’s guide to optimizing for generative AI features states that AI Overviews and AI Mode are “rooted in our core Search ranking and quality systems,” and that standard SEO best practices continue to apply (Google Search Central). The same guide punctures the growing folklore around AI search: there is no special schema.org markup you need to add, no llms.txt file required, no content “chunking” trick. Structured data remains useful for rich results — it is simply not a secret door into AI answers. That matters because a small industry is already forming around selling businesses special AI files and AI-specific rewrites. Google’s position is plainer: be indexed, be crawlable, publish content with a genuine point of view and firsthand experience, and don’t manufacture inauthentic mentions — the guide names that one specifically as not helpful. In other words, the foundation didn’t change. The answer layer on top of it did.

What the brand data shows

So if not a score, what correlates with actually showing up? Ahrefs analyzed 75,000 brands (all with Domain Rating above 40, so established businesses, not startups) against millions of AI Overview responses. The strongest correlation with AI Overview brand visibility was branded web mentions — 0.664 — mentions of the brand name anywhere across the web, linked or not. Domain Rating came in at 0.326. Raw backlink count: 0.218 (Ahrefs, 75K-brand AI Overview study). Read that carefully, because the study’s own authors insist on it: correlation is not causation. Well-known brands naturally attract mentions, searches, and citations all at once. But the direction is hard to ignore. What other people say about your business, in plain text, across the web, tracks your AI visibility roughly twice as strongly as the link metrics the SEO industry has sold for twenty years. And 26% of the brands in that study had zero mentions in AI Overviews at all. Invisible isn’t a ranking problem. It’s an evidence problem.

Ranking helps. It doesn’t finish the job.

Rankings still matter — seoClarity’s analysis of 362,000 AI Overviews (5.1 million citations) found that a URL in organic position 1 appears in the AI Overview about 43% of the time, with the likelihood sliding to 7% by position 20 (seoClarity research). Being near the top is a real edge. But here’s the finding I keep coming back to: 44% of AI Overview citations came from URLs outside the top 20 organic results. Nearly half the sources AI chose were not on the first two pages. Read the two studies together and a workable picture emerges: traditional visibility makes you competitive, and then something else — clarity, structure, topical fit, how extractable your actual answer is — decides whether you get picked for a specific question. Across those same 362,000 AI Overviews, 94% showed at least one citation overlapping the top 20 organic results — meaning almost every answer leans on established results, while still reaching beyond them.

The three outcomes nobody separates

This is where most businesses misread their own reports. A citation (your URL shown as a source), a mention (your brand named in the answer text), and a recommendation (the AI telling the asker to choose you) are three different outcomes, earned three different ways. A tool that blends them into one “AI visibility score” is hiding the only information you could actually act on. When I excavate a business’s evidence, these get measured separately — because the fix for “never cited” is not the fix for “cited but never recommended.” One useful lens, from WEB20 Ranker’s working model: authority gives a system a reason to consider you, extractability makes your information easy to retrieve, and corroboration is the independent evidence that reinforces both. It’s their framework, not a law of physics — but as a diagnostic it matches what I see in client work. The business with twelve years of expertise buried in walls of text has an extractability problem. The business with 600 polished AI-written articles and no reputation has an authority problem. Same symptom — AI ignores them — opposite prescriptions.

What this looks like in real life

In my own client work, this is measurable, not theoretical. Working with one established local retailer, we tracked how AI systems described and surfaced the business over time: visibility in ChatGPT rose from 10% to 20%, Gemini from 14% to 20%, and — the number I trust most — three or more walk-in customers came in naming ChatGPT as how they found the shop. No special AI file did that. Evidence did: clear information, consistent presence, and a business AI systems could actually understand and corroborate. That’s the work I call Business Archaeology — but only after the problem has a name. Years of expertise scattered across old campaigns, files, and other people’s websites never becomes reusable evidence on its own. Somebody has to excavate it, verify it, and put it where both humans and machines can find it.

Do this this week

Pick the five questions your best customers actually ask — the ones you’d want AI to answer with your business attached. Ask them, verbatim, in ChatGPT, Gemini, and Google’s AI Overview. Write down, for each one: Are you cited? Mentioned? Recommended? Or absent? That thirty-minute exercise will tell you more about your AI authority than any score a dashboard sells you — because authority in AI search isn’t a number someone assigns you. It’s the evidence the web holds about you, and evidence can be built, starting this week. If you want the structured version of that excavation, the AI Visibility Excavation Workbook ($27) walks you through auditing what AI knows, misunderstands, and can’t verify about your business. And the weekly newsletter at momentivemedia.kit.com is where I publish what the data actually says as this keeps moving. No gauge. Just evidence.

Frequently Asked Questions

What is an “AI visibility score,” and why doesn’t Google publish one?
It’s a third-party metric that blends different signals — citations, mentions, backlinks — into a single number. Google’s own guide to optimizing for AI features doesn’t define or endorse any single authority score; it talks about being indexed, crawlable, and demonstrating real expertise instead.

What actually correlates with showing up in AI Overviews?
Per Ahrefs’ study of 75,000 brands, branded web mentions correlate most strongly with AI Overview visibility (0.664), well ahead of Domain Rating (0.326) and raw backlink count (0.218).

Does ranking #1 on Google guarantee an AI Overview citation?
No. seoClarity’s analysis of 362,000 AI Overviews found a page ranking #1 appears in the AI Overview about 43% of the time, falling to roughly 7% by position 20 — and 44% of AI Overview citations came from outside the top 20 organic results entirely.

What’s the difference between a citation, a mention, and a recommendation?
A citation is your URL shown as a source. A mention is your brand named in the answer text. A recommendation is the AI telling the person to choose you. Each is earned differently, which is why blending them into one score hides what you’re actually missing.

How can I check my own AI visibility without a dashboard?
Ask the five questions your best customers actually ask, verbatim, in ChatGPT, Gemini, and Google’s AI Overview. Note for each one whether you’re cited, mentioned, recommended, or absent.

Sources

• Google — Optimizing for AI features on Search
• Ahrefs — An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)
• seoClarity — The Overlap Between AI Overviews and Organic Rankings

Frequently Asked Questions

What is an “AI visibility score,” and why doesn’t Google publish one?
It’s a third-party metric that blends different signals — citations, mentions, backlinks — into a single number. Google’s own guide to optimizing for AI features doesn’t define or endorse any single authority score; it talks about being indexed, crawlable, and demonstrating real expertise instead.

What actually correlates with showing up in AI Overviews?
Per Ahrefs’ study of 75,000 brands, branded web mentions correlate most strongly with AI Overview visibility (0.664), well ahead of Domain Rating (0.326) and raw backlink count (0.218).

Does ranking #1 on Google guarantee an AI Overview citation?
No. seoClarity’s analysis of 362,000 AI Overviews found a page ranking #1 appears in the AI Overview about 43% of the time, falling to roughly 7% by position 20 — and 44% of AI Overview citations came from outside the top 20 organic results entirely.

What’s the difference between a citation, a mention, and a recommendation?
A citation is your URL shown as a source. A mention is your brand named in the answer text. A recommendation is the AI telling the person to choose you. Each is earned differently, which is why blending them into one score hides what you’re actually missing.

How can I check my own AI visibility without a dashboard?
Ask the five questions your best customers actually ask, verbatim, in ChatGPT, Gemini, and Google’s AI Overview. Note for each one whether you’re cited, mentioned, recommended, or absent.

Sources

• Google — Optimizing for AI features on Search
• Ahrefs — An Analysis of AI Overview Brand Visibility Factors (75K Brands Studied)
• seoClarity — The Overlap Between AI Overviews and Organic Rankings

— Christina Rae Blackmon, Momentive Media Strategy beneath the surface. Visibility built on evidence.