Why Doesn't AI Mention My Brand? How AI Picks Sources

If you've asked ChatGPT or Perplexity for a recommendation in your category and your brand wasn't in the answer, you already know the problem. You might rank well in Google. Your website might be polished. Your reviews might be strong. None of that guarantees AI will mention you when a buyer asks for help.

The question most businesses start with is "how do I get my brand to show up in AI search?" That's the right starting point. The harder question is whether AI treats your brand as a recommendation or ignores it entirely. LLM Influence Score measures that distinction.

Google Rankings and AI Citations Are No Longer Connected

A year ago, you could reasonably assume that ranking well in Google meant you'd also appear in AI search answers. That assumption broke. Only 19 to 38% of AI Overview citations now come from Google's top 10 organic results for the same query. Across ChatGPT, Perplexity, and Google AI Overviews, only 14% of the most-cited sources are shared between all three platforms.

Your Google rankings tell you how Google's traditional algorithm sees you. They tell you almost nothing about whether ChatGPT will name you when someone asks "what's the best option for my situation."

These are now separate channels with separate mechanics. Winning one does not mean you're winning the other, and most businesses have no way to measure the second one.

Metric	ChatGPT	Perplexity	Google AI Overviews
How it finds sources	Searches Bing's index when triggered	Searches its own live index on nearly every query	Draws from Google's search index
How often it answers from memory	~60% of responses	Rarely; retrieval-first by design	Moderate; Google's Knowledge Graph acts as a memory layer
Most-cited source type	Wikipedia (7.8% of all citations)	Reddit (6.6%)	YouTube (9.4-10.7%)
Freshness preference	Strong; favors recently published content	Strong; live index updates continuously	Weak; cited content averages ~4 years old
Overlap with Google rankings	87% overlap with Bing top results	Not benchmarked	19-38% overlap with Google top 10

Metric

How it finds sources

ChatGPT

Searches Bing's index when triggered

Perplexity

Searches its own live index on nearly every query

Google AI Overviews

Draws from Google's search index

Metric

How often it answers from memory

ChatGPT

~60% of responses

Perplexity

Rarely; retrieval-first by design

Google AI Overviews

Moderate; Google's Knowledge Graph acts as a memory layer

Metric

Most-cited source type

ChatGPT

Wikipedia (7.8% of all citations)

Perplexity

Reddit (6.6%)

Google AI Overviews

YouTube (9.4-10.7%)

Metric

Freshness preference

ChatGPT

Strong; favors recently published content

Perplexity

Strong; live index updates continuously

Google AI Overviews

Weak; cited content averages ~4 years old

Metric

Overlap with Google rankings

ChatGPT

87% overlap with Bing top results

Perplexity

Not benchmarked

Google AI Overviews

19-38% overlap with Google top 10

For the technical breakdown of why these channels diverged, see Retrieval Layer and Parametric Layer: The Two Systems That Determine AI Search Visibility.

Most Brands That Think They're Invisible Are Actually Being Found and Ignored

Your content is probably already being found by AI search platforms. AI sees you, evaluates you, and passes you over in favor of something else.

Research from early 2026 found that only about 15% of the pages AI retrieves during a search actually make it into the final answer. The rest are evaluated and discarded. Your content enters the room, gets looked at, and gets passed over.

This is the gap between retrieval and citation, and it's where most brands lose. A monitoring tool that checks "did AI find my page?" gives you a green light while the actual outcome is a closed door. Your page was found. It wasn't used. The monitoring tool can't tell the difference.

For the full mechanics of how this pipeline works, see Why AI Cites Your Content and Recommends Your Competitor.

The Brands That Keep Appearing Have Something You Can Build

When business owners describe this problem, they often say it feels like AI has a shortlist of approved brands. That instinct is closer to correct than most practitioners acknowledge.

AI search platforms operate with two knowledge sources. One is what they retrieve in real time from the web. The other is what they already know from training data: a built-in memory of which brands exist, what they do, and how they relate to each other. Approximately 60% of ChatGPT responses draw primarily from this memory layer, not from live web retrieval.

The brands that keep appearing in AI answers have strong signals in both layers. AI's memory already includes them as a credible option in the category, and their current content is structured in a way that makes it easy to cite. That combination is what produces consistent visibility. Without both, you're competing with one hand tied.

For more on how AI's memory layer works and why it matters, see Parametric vs. Retrieval Knowledge: When Models Answer From Memory.

Your Page Can Be Cited Without Your Brand Being Mentioned

There's a third problem that most measurement tools miss entirely. Your URL can appear in AI's source list while your brand name never appears in the actual answer the user reads.

Research across more than 500,000 AI responses found this pattern at scale: the URL is cited as a reference, but the response text names a competitor as the recommendation. The monitoring dashboard says you were cited. The buyer never saw your name.

This is why counting citations is not the same as measuring influence. A citation that doesn't produce a brand mention has no business value. The user reads the answer, sees the recommended brands, and may never look at the source list. You're technically present and practically invisible.

For the full breakdown of this phenomenon and what drives it, see Why AI Cites Your Content and Recommends Your Competitor.

What LLM Influence Score Actually Measures

LLM Influence Score exists because citation counting doesn't capture what matters for business outcomes. It measures four things:

Citation frequency tracks how often your content appears in the source list across platforms. This is what most monitoring tools measure. It's the baseline, not the goal.

Answer inclusion tracks whether your brand is named in the response text. This is the difference between "your URL was in the footnotes" and "ChatGPT mentioned you by name when answering the question."

Recommendation status tracks whether AI positions your brand as a viable solution. Being mentioned as background context ("companies like X operate in this space") is different from being recommended ("X is a strong option for this use case"). The business impact is entirely different.

Description accuracy tracks whether AI describes your brand correctly. A confident, wrong description can be worse than no mention at all. If AI tells a buyer you offer something you don't, or positions you in the wrong category, that's active damage, not just missed visibility.

Together, these four components measure what citation counting can't: whether AI is actually helping buyers find you, or whether your presence in AI search is a technical footnote with no business value.

Why Prompt-Based Testing Doesn't Give You a Reliable Score

A common approach to measuring AI visibility is to run a set of test prompts across platforms and log what comes back. The problem is that AI search responses are not stable. The same question asked twice can produce different results because AI decomposes each question into sub-queries shaped by conversation context, session history, and real-time assessment of what the question needs.

A prompt audit captures one snapshot of one response at one moment. It doesn't tell you what a different buyer, asking a slightly different version of the same question, will see tomorrow.

LLM Influence Score works from the other direction. Instead of simulating buyer questions and hoping the snapshot is representative, it measures the structural inputs that drive citation outcomes: whether AI can clearly identify who you are and what you do, whether AI's built-in memory describes your brand accurately, how well your content is structured so AI can quote specific parts of it, and how strong your reputation is across the review sites, directories, and publications AI checks in your category.

These are the upstream causes. Citations are the downstream effect. Measuring causes tells you what to fix. Measuring effects only tells you what happened once.

For how content structure specifically affects whether AI can use your pages as citation sources, see AI Systems Are Not Googlebot. They Answer Questions With Chunks, Not Entire Pages.

Frequently Asked Questions

Why doesn't ChatGPT mention my brand when someone asks about my category?

ChatGPT draws approximately 60% of its responses from its built-in memory, which was formed during training. If your brand wasn't well-represented in the training data through Wikipedia, widely-cited publications, or consistent third-party coverage, the model has no memory of you. It will only mention you if its live search retrieves your content and your content survives reranking and citation selection. LLM Influence Score measures both layers: whether AI's memory includes you and whether your content is structured to be cited when retrieval does happen.

We rank on page one of Google. Why are we still invisible in AI search?

Google rankings and AI citations are now separate outcomes. Only 19 to 38% of AI Overview citations come from Google's top 10 organic results, and only 14% of the most-cited sources are shared across ChatGPT, Perplexity, and Google AI Overviews. Strong Google rankings mean Google's traditional algorithm values your page. AI search platforms evaluate content at the passage level using different signals, including entity clarity, semantic density, and whether your content answers the specific sub-questions the model generates when it decomposes a query. These are different channels that require separate measurement and separate optimization.

How is LLM Influence Score different from tracking AI citations?

Citation tracking measures whether your URL appears in AI's source list. LLM Influence Score measures four outcomes: citation frequency (are you in the source list), answer inclusion (is your brand named in the response text), recommendation status (does AI position you as a viable solution), and description accuracy (does AI describe you correctly). Research across more than 500,000 AI responses found that a URL can appear as a citation while the brand is never mentioned in the answer the user reads. Citation presence alone has no business value if the buyer never sees your name.