AI visibility: your score hides two different problems

Ford appears in all four major AI engines, but its site ranks beyond position 700 among cited sources. If you only look at your AI visibility "score," you won't see this problem, and you risk spending in the wrong place.

Is my AI visibility score reliable, or is it misleading me?

A single AI visibility score blends two very different metrics: how often your brand is named in answers and how often your site is used as a source. Confusing them leads you to invest in the wrong place.

Many monitoring tools return just one number: your AI visibility. It seems convenient, but it hides a distinction that changes everything.

Being mentioned in an AI answer means the model knows your brand and names it in the text. Being cited as a source means your site appears among the reference links used to build that answer. They are two very different positions, with different causes and different remedies.

The Semrush 2026 AI Visibility Index, based on 126 million prompts analyzed across ChatGPT, Gemini, Google AI Mode and AI Overviews, measured this gap across 22 categories and two markets (US and UK). The result is clear: on average, fewer than half of the 100 most-mentioned brands also appear among the 100 most-cited domains, every month.

What did the Semrush research really find about the best-known brands?

Global brands like Ford, Disney and Toyota show up in all AI answers but fall outside the top 500 cited sources. Ford is the most striking case: mentioned everywhere, almost invisible as a source.

The Semrush 2026 AI Visibility Index analyzed concrete cases that show just how extreme the gap can be. In July 2025, Ford ranked in the top 100 for mentions on all four AI engines. But as a cited domain? Position 1,055 on ChatGPT, 1,146 on Gemini, 1,320 on AI Mode, 748 on AI Overviews.

The opposite phenomenon exists just as much. Investing.com, in the US Finance sector on ChatGPT, ranked seventh among cited domains but fiftieth among mentioned brands: a great source provider, with little presence in direct answers.

Chevrolet, Disney, Honda, Marvel, Nintendo and Toyota share Ford's situation: top 100 for mentions on all four engines, outside the top 500 for citations every month. They're very strong brands in AI's perception, but their sites aren't used as proof.

Do ChatGPT and Gemini work the same way for my brand's visibility?

No: the two engines behave in almost opposite ways. ChatGPT cites a lot of sources but names few brands; Gemini names a lot of brands but cites very few sources. Optimizing for one doesn't guarantee results on the other.

This is the part that surprises marketing managers the most. ChatGPT and Gemini aren't interchangeable: they follow very different answer logic, and a company that performs well on one can be almost invisible on the other.

According to Semrush data, ChatGPT cites an average of 15 sources per answer, while Gemini cites only 3 on average. On Gemini, the overlap between mentioned brands and cited domains drops as low as 30%.

An earlier Semrush study of 3,981 domains measured the opposite behavior even more sharply: ChatGPT cites the source in 87% of cases but names the brand in only 20.7% of answers. Gemini names brands in 83.7% of answers but cites the source in only 21.4% of cases. These are two engines with almost opposite priorities, as also explained in our analysis of ChatGPT and primary sources.

Google's AI Overviews also behaves differently from the other three: its brand mentions overlap less with those of the other engines, and its cited sources have the least overlap with ChatGPT. A good result on ChatGPT tells you nothing about where you stand in AI Overviews.

To understand how citation logic works in AI Overviews, it's also worth reading our analysis of AI Overviews and brand visibility.

What changes in practice for those managing communications at an Italian company?

The type of gap determines who should work on it: a mention gap is a brand and PR problem, while a citation gap is a problem of technical content and site authority. Confusing them means giving the job to the wrong person.

The practical consequence is simple but often overlooked: the team that needs to fix the problem depends on the type of gap. If your brand isn't mentioned, the job is brand awareness: articles in industry publications, a presence on Wikipedia, mentions in forums and communities. If your site isn't cited as a source, the job is technical and editorial: authoritative content, schema markup (code that helps AI engines understand what a page is about), links from trusted sites.

Using the same tactic for both problems is the most common waste. As our analysis of how to earn AI citations and measure visibility also explains, the distinction between mentions and citations is the starting point of any sensible strategy.

Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?

The type of gap varies widely by industry and company type. Here are three examples showing how to diagnose it and where to step in.

Illustrative examples: numbers are meant to show the reasoning, they are not real customer data.

In the metalworking case, SHM Studio's experience with B2B companies in the industrial sector shows a recurring pattern: the site gets technically optimized and climbs in AI citations, but the brand stays barely recognizable in the answers because there are no mentions in industry publications and communities. SEO work alone doesn't close that gap.

Do brand mentions matter more than links for AI visibility?

Yes, and the gap is clear. Brand mentions across the web correlate with AI Overview visibility almost three times more than traditional backlinks.

Anyone who has worked in SEO for years knows that backlinks, the links other sites point to yours, were the most important signal for Google for decades. With AI engines, the hierarchy changes.

According to an Ahrefs study of 75,000 brands (August 2025), brand mentions across the web correlate with AI Overview visibility with a coefficient of 0.664, while backlinks correlate at only 0.218. Mentions are roughly three times more predictive than links for AI visibility.

This doesn't mean links no longer matter. It means a strategy that focuses only on links and ignores unlinked mentions leaves out a big part of the signal AI engines use to decide whom to trust.

The topic of external mentions and their weight in AI visibility is also explored in our article on brand context and AI, which explains how engines build their perception of a brand from what third-party sites say.

Does my industry change the rules of the game?

A lot. In some industries the sources cited by AI engines are stable and predictable; in others they change constantly. Knowing which category you're in changes the strategy you should adopt.

The Semrush 2026 AI Visibility Index found that differences between industry categories outweigh differences between AI platforms in mention-citation overlap. The industry matters more than the engine.

In US Real Estate, for example, the same domains (Apartments.com, Realtor.com, Redfin, Zillow) show up as sources on all four engines every month. Anyone operating in that sector already knows who they have to compete with for citations. In the US Industrial sector, on the other hand, there is no stable core of recurring sources: the landscape changes every month.

For an Italian company operating in a fragmented sector, this means a generic category benchmark tells you little. You need to start from the specific questions your customers ask AI engines, not from an industry average.

Why it's being discussed now: are metrics really changing in the market?

Yes: 86% of companies are already changing their measurement metrics to adapt to AI search. The problem isn't in the future. It's here now.

The pressure to change approach doesn't come only from technical data. It comes from the market. According to an IAB study from September 2026, 44% of advertisers name adapting to AI search as their main media investment challenge, and 86% say they are changing their measurement metrics.

In Europe, according to IAB Europe (September 2025), 85% of companies already use AI tools for marketing, mainly for targeting (64%) and content generation (61%). AI is no longer an experiment: it's already part of the process. The question is whether the metrics you use to measure results are still the right ones.

The way citations work in AI Overviews is also more complex than it looks: according to a Pew Research analysis (March 2025), 88% of AI Overviews cite three or more sources, and Wikipedia, YouTube and Reddit together account for 15% of all sources cited. That means focusing only on your own website isn't enough.

What mistakes are made most often when trying to improve AI visibility?

The most common mistake is using the same tactic for different problems: producing more content when the gap is in external mentions, or doing PR when the gap is in the site's technical authority.

Here are the most common mistakes we see in companies that start working on AI visibility.

One underrated mistake involves Bing and its analytics tools: as explained in the article on Bing Topics Portfolio and AI visibility, even the engines people overlook offer valuable diagnostic data that many companies ignore completely.

How to get started: the steps to diagnose your gap

Always start by measuring mentions and citations separately, using a fixed set of questions your customers really ask AI engines. Only then do you choose the tactic.

The operating principle suggested by the Semrush 2026 AI Visibility Index is simple: diagnose the type of gap first, then assign the work to the team that can actually solve it. Here's how to do it in practice.

How can you tell if your actions are really working?

Keep AI visibility metrics separate from business results. More mentions don't automatically mean more sales: track both, but don't confuse them.

The Semrush 2026 AI Visibility Index is explicit on this point: the study doesn't establish that a change in AI visibility causes a commercial result. The link exists, but it isn't automatic. Tracking both is necessary, but keeping them separate is just as important.

To dig deeper into building an AI visibility measurement system that truly supports decisions, the article on LinkedIn and AI citations for B2B brands offers a practical view of how signals from different platforms add up.

On the subject of sentiment, it's also worth keeping an eye on BrightEdge data (March 2026): AI Overviews is 44% more likely than ChatGPT to generate negative sentiment about brands, while ChatGPT is 13 times more likely to do so in the stages close to purchase. It's not just a question of visibility, but of tone too.

So, is it worth tackling this problem now for your company?

It's worth it if your industry is already covered by stable sources on AI Overviews and ChatGPT, and you want to understand where you stand against them. Or you already have solid classic SEO and want to know whether your site is being used as an AI source.

Somewhere in between you're starting to measure AI visibility but still use a single aggregate score. The first step is to separate mentions and citations for each engine that matters in your industry.

Not yet if your brand doesn't have a structured digital presence yet: before working on AI visibility, strengthen your content, website and basic editorial presence. AI visibility amplifies what exists. It doesn't create something from nothing.

How we see it at SHM

For us, the difference between a mention and a citation isn't a technical subtlety: it's the difference between knowing what to do and wasting budget. We believe simplifying digital is the highest form of expertise — and here, simplifying means giving the problem a precise name before choosing the solution.

Working with B2B companies in the industrial sector, we've often seen this pattern: the site climbs in AI citations thanks to solid technical SEO, but the brand stays barely recognizable in answers because there are no mentions in industry publications and communities. These are two different jobs, to be done in parallel, not in sequence.

Our approach always starts with diagnosis: we measure mentions and citations separately for the AI engines relevant to your industry, inspect the sources being cited instead of you, and only then decide where to act. If the gap is in authority as a source, we work on SEO for AI and content architecture. If the gap is in brand awareness, the work involves editorial strategy and presence on third-party platforms. If you want the full picture, the starting point is our area dedicated to SEO and AI visibility.

If you want to understand where your brand really stands on the AI visibility map, get in touch: we start with diagnosis, not with the solution.

Frequently Asked Questions

The problem isn't how visible you are in AI. It's knowing whether you're working on the right cause.