As conversational artificial intelligence becomes an integral channel for discovery, enterprise brands and digital marketers are racing to understand how Large Language Models (LLMs) like ChatGPT perceive market leadership. For years, Search Engine Optimization (SEO) relied on predictable ranking factors: high-quality backlinks, domain authority, keyword density, and technical site performance. However, as user search behavior shifts toward generative AI interfaces, traditional organic search benchmarks are failing to predict which brands win market share inside AI-generated answers.
A comprehensive study conducted by Semrush in partnership with Kevin Indig, founder of Growth Memo, offers a detailed look into how ChatGPT handles brand recommendations across enterprise sectors. The findings reveal a stark reality: true topic dominance in ChatGPT is exceptionally rare, and standard SEO metrics offer almost no guarantee that an AI engine will consistently favor a specific brand across the entire buyer journey.
Understanding ChatGPT Topic Ownership: The Benchmark Data
To measure brand dominance in AI search responses, Semrush and Kevin Indig analyzed a massive dataset spanning six months, tracking brand visibility in ChatGPT from January through June 2026. The research examined 1,094 U.S. commercial topic categories using the Semrush AI Visibility Toolkit. The scope of the study was vast, evaluating:
- More than 50,000 distinct brands
- 220,000 unique web domains
- 600,000 external citations
- 220,000 distinct source URLs
Rather than evaluating brand performance based on a single search prompt, the study analyzed category strength across five distinct buyer intent stages. For each of the 1,094 categories, researchers queried ChatGPT with five prompts tied directly to standard consumer purchase decisions:
- Definitions: What is the service or product category?
- Comparisons: How do leading solutions evaluate against one another?
- Alternatives: What options exist when considering major market providers?
- Use Cases: Which solution is best suited for specific organizational or consumer scenarios?
- Buying Decisions: Which provider should a customer ultimately choose?
By mapping responses across these five touchpoints, researchers could determine whether ChatGPT consistently recognized a single brand as an authoritative category leader.
The Rarity of Brand Dominance in Generative Responses
The central discovery of the study is that only 15.2% of ChatGPT topic categories had a clear brand owner. That leaves nearly 85% of evaluated business categories without a dominant brand presence in generative AI answers. In the overwhelming majority of topics, appearing in an AI response for one prompt did not mean the brand maintained visibility when the question was reframed or approached from another angle within the buyer journey.
To establish a clear definition of category ownership, Semrush instituted a rigorous methodology. A brand was classified as a “Category Owner” only if it met all of the following criteria:
- Captured the highest overall share of brand mentions within the category.
- Appeared in at least four out of the five related buyer intent prompts.
- Maintained a lead of at least 5 percentage points over the nearest competing runner-up brand.
Out of the 1,094 commercial categories evaluated, only 166 categories met this standard. The remaining spectrum of commercial topics fell into two distinctly fragmented tiers:
Emerging Leaders (31.2%): In 341 categories, a leading brand surfaced in at least three of the five prompt types, but failed to establish the required 5 percentage point lead over the second-place brand. These categories reflect active competition where no single entity has locked down generative mindshare.
Unsettled Categories (53.7%): In 587 categories—representing more than half of the entire study—no single brand appeared in even three of the five prompts. In these sectors, ChatGPT provided highly varied, fragmented, and inconsistent brand recommendations depending on how the prompt was worded.
The Search Volume Paradox: High Demand Means Higher Fragmentation
A intuitive assumption might suggest that high-demand topics—where commercial competition is fiercest—would produce established, dominant market winners. However, the Semrush dataset revealed the exact opposite phenomenon.
When dividing the 1,094 categories into two equal halves based on AI search volume, a stark contrast emerged. The top half of categories accounted for an overwhelming 98% of total AI search demand in the sample. Yet within this high-demand group, only 11.3% of topics had a clear brand owner.
In contrast, the lower-demand half of categories—representing just 2% of total search volume—saw clear topic ownership reach 19.0%.
This distribution highlights an essential dynamic in AI search: high-demand commercial spaces suffer from extreme content competition and diversified web coverage. Because thousands of publishers, review sites, and competitors produce content around high-volume topics, ChatGPT’s underlying models digest millions of conflicting signals. As a result, generative answers synthesize a broader mix of brands, preventing any single company from dominating the topic across multiple query variations.
Why Traditional SEO Metrics Fail to Predict AI Topic Ownership
For decades, digital marketing executives relied on metrics like Domain Authority, backlink volume, and organic keyword rankings to forecast search dominance. However, Semrush found that traditional SEO strength offers surprisingly weak predictive value when trying to determine who owns a topic inside ChatGPT.
When comparing clear category owners against their runner-up competitors across primary SEO performance indicators, the correlation was remarkably inconsistent:
- Branded Search Volume: Category owners had higher branded search volume than their runner-up in 55.7% of comparisons. This was the only traditional SEO metric in the study that reached statistical significance.
- Organic Search Traffic: Category owners possessed higher overall organic traffic in only 48.4% of comparisons—meaning the runner-up had higher organic web traffic more than half the time.
- Authority Score: Category owners held a higher Semrush Authority Score in just 52.5% of cases, essentially performing no better than a coin flip.
“Traditional SEO metrics aren’t enough to explain who owns a topic. While they play their role, there’s more to it,” noted Kevin Indig, founder of Growth Memo.
The failure of standard SEO metrics to predict AI visibility stems from the core architecture of LLMs. Traditional search algorithms rely heavily on link graphs and on-page crawl data to index and rank web pages. In contrast, generative models like ChatGPT rely on probabilistic natural language modeling, entity co-occurrence across broad training corpora, and Retrieval-Augmented Generation (RAG) frameworks. A brand with a high domain authority may rank #1 on Google, but if the broader web narrative does not consistently pair that brand with specific industry solutions, an LLM will not reliably recommend it in conversational prompts.
Mentions vs. Citations: The Generative Disconnect
One of the most critical operational insights from the study is the sharp distinction between brand mentions in generative text and source citations in footnote links.
When tracking topic ownership, Semrush measured brand mentions within the synthesized conversational text generated by ChatGPT, rather than evaluating the reference links displayed at the bottom of the response. The research revealed that text mentions and link citations rarely aligned.
Across the entire dataset, only 21% of the most-cited web domains in a category were also the most-mentioned brand in the answer text.
This disconnect reveals a crucial nuance for modern digital publishing and enterprise PR. An organization may publish high-ranking, authoritative content that ChatGPT frequently retrieves and cites in its reference panel. However, if that content references third-party tools, market research, or alternative platforms, ChatGPT may output answer text recommending competitor brands while simply using the publisher’s domain as an informational citation link. Being used as a data source by an LLM is not the same as being recommended as the commercial solution.
Rank Stability: Defending the AI Crown
While establishing topic ownership in ChatGPT is difficult, the study indicates that once a brand successfully locks down a clear lead, that dominance is surprisingly durable.
In month-over-month comparisons tracking brand performance from January through June 2026, clear category owners retained their top position in 90.4% of instances. Once an LLM forms a dense, positive entity association across multiple buyer journey prompts, the association tends to persist across model runs and minor dataset updates.
Conversely, non-dominant positions proved highly volatile. In emerging and unsettled categories, market leadership changed constantly. Across 5,470 month-over-month comparisons in non-owned categories, the leading brand switched places 1,950 times.
The stability of a brand’s lead directly correlated with the size of its advantage over competitors:
- When a leading brand eventually lost its top spot in subsequent months, its average prior lead over the runner-up was just 1.3 percentage points.
- When a leading brand successfully defended its top spot month-over-month, its average lead was 2.9 percentage points—more than twice as large.
While the study noted that the data does not explicitly explain the technical mechanisms causing these monthly shifts, it clearly demonstrates that narrow leads in generative search are inherently unstable, whereas wide brand margins create a defensible moat.
Strategic Takeaways for Generative Engine Optimization (GEO)
The findings from the Semrush and Kevin Indig study mark a decisive shift in how brands must approach search visibility moving forward. As consumer reliance on ChatGPT and generative answer engines grows, search marketers must pivot from single-keyword rankings to topic-level entity optimization.
To build defensible brand visibility within generative AI ecosystems, digital publishing teams and brand strategists should adopt several key principles:
1. Measure Visibility at the Topic Level, Not the Single Prompt: Ranking for a isolated prompt like “best CRM software” offers a false sense of security. Because 85% of categories lack clear ownership across related buyer questions, brands must audit their presence across the entire buyer journey—including comparisons, use cases, and alternative queries.
2. Prioritize Brand Building and Demand Generation: Because branded search volume was the only traditional SEO metric with a statistically significant correlation to ChatGPT topic ownership (55.7%), off-page brand demand remains crucial. Broad media exposure, PR coverage, and organic consumer brand search signals actively reinforce entity recognition within generative training data.
3. Align Citations with Entity Alignment: Earning citations from high-authority sites is no longer enough if those citations do not explicitly frame your brand as a category leader. PR and content efforts must focus on co-occurrence—ensuring that whenever your industry category is discussed across authoritative publications, your brand is explicitly named alongside key problem-solving use cases.
4. Monitor Entity Sentiment Beyond Organic Rankings: Traditional rank trackers that only record blue link placements are blind to generative dynamics. Utilizing modern toolkits like the Semrush AI Visibility Toolkit allows enterprise teams to measure actual text mentions, conversational voice share, and competitive lead margins within LLM outputs.
As detailed in the full analysis published in the Semrush AI Visibility Study, winning visibility in the generative age requires moving past simple SEO tricks. Brands that wish to own their category in ChatGPT must build pervasive, consistent, and cross-platform authority that proves to AI models—and consumers alike—that they are the definitive leader in their space.