Key Statistics at a Glance
Source: AISearchStackHub curated research study, January–April 2026. 527 brands, 4 LLM engines, 24 queries per brand. See Methodology section below.
Live Platform Data (updates with every scan)
Data source: AISearchStackHub proprietary scan database. Anonymized aggregate only — no PII or brand names. Same AIS Index scoring methodology as the curated research study above.
Executive Summary
The shift from traditional search to AI-native information retrieval is not a future event — it is happening now. As of early 2026, an estimated 30–40% of commercial information queries in the United States are being answered at least in part by a large language model rather than a traditional SERP. ChatGPT alone reports over 100 million weekly active users. Perplexity is processing over 10 million queries per day. Claude and Gemini are embedded in productivity tools used by hundreds of millions of people.
Yet most brands are effectively invisible inside these systems. Our analysis of over 500 brands across four major verticals — B2B SaaS, ecommerce, fintech, and healthcare — finds that 74% score below 40 on the AISearchStackHub AIS Index, a composite measure of LLM brand visibility. The average score across all brands is 31 out of 100. Only the top 10% of brands score above 68.
This report documents the current state of AI search visibility, explains what separates high-scoring brands from low-scoring ones, compares citation behavior across the four leading LLMs, and provides vertical-specific benchmarks for B2B SaaS, ecommerce, fintech, and healthcare. It is intended as a reference document for marketers, SEO professionals, and brand strategists making decisions about AI search optimization in 2026.
The methodology is described in detail in Section 2. All data was collected between January and April 2026 using AISearchStackHub's automated scanning infrastructure, which issues standardized queries to each LLM and scores responses on four dimensions: Visibility, Authority, Sentiment, and Advantage.
Methodology
Brand Selection
We analyzed 527 brands selected to be representative of their respective verticals. Selection criteria: the brand must operate a public-facing website with commercial offerings, must have existed for at least 24 months as of January 2026, and must have some form of publicly available documentation, pricing, or press coverage. We excluded brands with fewer than 10 employees and brands that had explicitly opted out of LLM training or web indexing.
The final sample comprised 147 B2B SaaS brands, 142 ecommerce brands, 118 fintech brands, and 120 healthcare brands.
Query Design
For each brand, we issued 24 standardized queries across four categories. These queries were designed to reflect how real users and buyers encounter brands in LLM responses — not direct name searches (which any brand can rank for), but category-level and problem-level queries where brand mention is organic.
- Category queries (6): "What are the best [category] tools?" — designed to test baseline category presence.
- Problem queries (6): "How do I solve [specific problem]?" — tests whether LLMs recommend the brand as a solution.
- Comparison queries (6): "Compare [brand] vs alternatives" and "What's better than [competitor in same space]?" — tests competitive visibility.
- Authority queries (6): "What do experts say about [domain]?" and "What's the evidence for [brand's core claim]?" — tests citation authority.
Queries were issued to ChatGPT (GPT-4o), Claude 3.5 Sonnet, Perplexity Pro, and Gemini 1.5 Pro. Each query was issued three times per engine with a 24-hour gap between repetitions to account for model stochasticity. Scores represent the average across three runs.
AIS Index Scoring Formula
The AIS Index is a composite score calculated from four sub-scores, each measuring a distinct dimension of LLM brand visibility:
- Visibility (V, weight 40%): How often does the brand appear in LLM responses at all? Measured as mention rate across all 24 queries × 3 repetitions × 4 engines.
- Authority (A, weight 30%): When the brand appears, is it cited as a primary recommendation, or mentioned peripherally? Primary mentions, first-position mentions, and explicit recommendation language all increase the Authority score.
- Sentiment (S, weight 20%): Is the brand mentioned positively, neutrally, or negatively? LLM responses are scored by a secondary analysis pass that classifies sentiment on a -2 to +2 scale, normalized to 0–100.
- Advantage (Ad, weight 10%): Does the LLM associate the brand with specific differentiated capabilities? Mentions of unique features, proprietary data, or distinctive methodology contribute positively.
Key Findings
Finding 1: The vast majority of brands are LLM-invisible
The headline finding is stark: 74% of brands in our study score below 40 on the AIS Index, meaning they rarely or never appear in LLM responses to category and problem queries. A score below 20 — earned by 31% of brands — indicates near-total invisibility: the brand is essentially absent from AI-mediated information environments.
A score of 40–60 indicates moderate visibility: the brand appears in some LLM responses, is occasionally cited as a primary recommendation, and has neutral-to-positive sentiment when mentioned. About 18% of brands fall in this range. A score above 60 indicates strong visibility — the brand appears consistently, is often recommended, and has positive sentiment. Only 8% of brands achieve this.
| AIS Score Range | Category | % of Brands |
|---|---|---|
| 0–19 | Near-invisible | 31% |
| 20–39 | Low visibility | 43% |
| 40–59 | Moderate visibility | 18% |
| 60–79 | High visibility | 6% |
| 80–100 | Market-defining | 2% |
Finding 2: Perplexity has the highest citation rate of any LLM
Perplexity's architecture — which performs real-time web search before generating responses — produces substantially different citation behavior from other LLMs. In our study, 87% of Perplexity responses to brand-adjacent queries cited at least one named source. The comparable figures were 68% for ChatGPT with browsing, 61% for Claude with web search, and 43% for Gemini 1.5 Pro.
This has a practical implication: brands that publish structured, crawlable content on high-authority domains have a disproportionately higher Perplexity visibility score than their overall AIS score. The average Perplexity sub-score across all 527 brands in our dataset is 44/100, compared to an overall AIS average of 31/100.
Finding 3: ChatGPT cites sources 2.3x more than Gemini
Among the four engines, ChatGPT and Gemini show the widest divergence in citation behavior. ChatGPT (GPT-4o with browsing) produces named source citations in 68% of responses. Gemini 1.5 Pro — despite access to Google's search index — produces named citations in only 30% of responses. Claude sits at 61%, Perplexity at 87%.
Gemini's lower citation rate is partly explained by its tendency toward synthesized responses that summarize information without attributing specific sources. This makes Gemini harder to influence through traditional AEO tactics and more dependent on training-time representation. Brands that were mentioned frequently in pre-training data (news coverage, Wikipedia, high-authority publications) score better on Gemini than brands that rely primarily on owned media.
Finding 4: B2B SaaS brands significantly outperform ecommerce
B2B SaaS brands average 38/100 on the AIS Index — the highest of the four verticals studied. Ecommerce brands average 24/100 — the lowest. The gap is explained by structural differences in how these categories are represented in LLM training data and in current web content.
B2B SaaS brands tend to publish more technical documentation, case studies, integration guides, and comparison content — exactly the type of structured, factual, reference-quality content that LLMs learn to cite. Ecommerce brands, by contrast, invest heavily in product pages and ad creative — content formats that LLMs rarely cite in response to research or recommendation queries.
Vertical Breakdown
The four verticals in this study show meaningfully different AIS Index distributions, reflecting both structural differences in content publishing behavior and differences in how LLMs were trained on each category's domain knowledge.
<\!-- SaaS -->B2B SaaS — Average AIS Score: 38/100
B2B SaaS is the strongest-performing vertical in our study. The top quartile of B2B SaaS brands scores above 61/100. The performance advantage stems from an inherent content advantage: SaaS companies must publish documentation, API references, onboarding guides, and integration catalogs to support their customers. This content is crawlable, structured, factual, and regularly updated — all characteristics that correlate with higher LLM citation rates.
SaaS brands that publish transparency data — pricing pages with exact figures, customer count statistics, and quantified outcome metrics ("reduce churn by 28%") — score significantly higher on Authority than those that use vague language. In our analysis, brands with at least three public case studies featuring quantified outcomes score on average 14 points higher than those without.
Within B2B SaaS, CRM tools and data analytics platforms score highest (averages of 44 and 42 respectively). Project management tools score lowest at 31, likely because the category is saturated with near-identical feature claims, making it difficult for LLMs to differentiate among options.
Ecommerce — Average AIS Score: 24/100
Ecommerce brands are the weakest performers in our dataset, with 53% scoring below 20/100. This is not primarily a content volume problem — most ecommerce brands publish substantial content. It is a content type problem: product descriptions, promotional copy, and ad-optimized blog posts are not the types of content that LLMs treat as authoritative sources.
Ecommerce brands that outperform their vertical average share a common trait: they publish research-style content about their product category. A supplement brand that publishes peer-reviewed summaries of ingredient research. A furniture brand that publishes guides on sustainable materials sourcing. A fashion brand that publishes detailed breakdowns of manufacturing quality standards. These content types anchor LLM associations between the brand and authoritative domain knowledge.
Top-quartile ecommerce brands score above 47/100, indicating that the gap between the best and worst performers is substantial and that the vertical is not structurally incapable of achieving high AIS scores — it just requires a different content strategy than what most ecommerce brands currently deploy.
Fintech — Average AIS Score: 33/100
Fintech brands occupy the middle of the vertical ranking, with an average score of 33/100 and a top quartile threshold of 56/100. Fintech benefits from the fact that LLMs are frequently queried for financial product comparisons and recommendations — "best business checking account," "cheapest way to send international transfers," "which robo-advisor has the lowest fees" — which creates natural visibility opportunities for well-positioned brands.
However, fintech brands face a unique constraint: LLMs tend to apply extra caution in financial contexts, often hedging recommendations and directing users to consult financial advisors. This reduces the Authority score for fintech brands relative to other verticals where LLMs are more comfortable making direct recommendations.
Fintech brands that publish fee transparency data, comparison tables, and user outcome data (average savings, return rates, approval rates) see Authority scores 11 points above vertical average. Regulatory disclosures that establish legitimacy — FDIC membership, SEC registration, state licensing — also correlate with higher LLM citation rates.
Healthcare — Average AIS Score: 29/100
Healthcare brands face the most significant structural challenges for LLM visibility. All four LLMs studied apply YMYL (Your Money or Your Life) caution in health-adjacent queries, defaulting to recommendations for qualified healthcare providers rather than specific brands. This suppresses Authority scores across the entire vertical.
Despite these constraints, top-quartile healthcare brands reach scores above 52/100, typically by establishing themselves as educational authorities rather than direct-to-consumer vendors. Healthcare brands that publish clinical study summaries, maintain medical advisory boards with cited credentials, and provide condition-level educational content score substantially higher than those focused primarily on product promotion.
The Visibility sub-score is less affected by YMYL caution than the Authority sub-score — healthcare brands appear in LLM responses, they just appear in supporting roles rather than as direct recommendations. AEO strategy for healthcare brands should prioritize Visibility and Sentiment over Authority, with Authority investment focused on establishing category credibility rather than driving direct recommendation.
| Vertical | Sample Size | Avg AIS Score | Top Quartile | % Below 20 |
|---|---|---|---|---|
| B2B SaaS | 147 | 38 | 61 | 19% |
| Fintech | 118 | 33 | 56 | 27% |
| Healthcare | 120 | 29 | 52 | 34% |
| Ecommerce | 142 | 24 | 47 | 53% |
Engine Comparison: ChatGPT vs Claude vs Perplexity vs Gemini
The four LLMs in our study behave differently in ways that matter substantially for brand visibility strategy. Understanding these differences allows brands to prioritize their AEO investments toward the engines that are most likely to drive awareness among their target audience.
- + High citation rate (68% of responses cite sources)
- + Strong for comparison and vs. queries
- + References review sites, news, official docs
- – Inconsistent real-time indexing
- – Citation rate varies widely by query type
- + Highest sentiment accuracy (nuanced brand description)
- + Strong on authority queries (cites research, studies)
- + Well-calibrated on YMYL topics
- – More conservative recommendation language
- – Lower mention rate on category/product queries
- + Highest citation rate (87% of responses)
- + Real-time web index — fastest propagation
- + Best for brands with fresh, structured content
- – Skewed toward high-DA domains
- – Newer brands disadvantaged
- + Strong for brands with Wikipedia/news presence
- + Excellent Knowledge Graph integration
- + Good coverage for established enterprise brands
- – Lowest citation rate (30% cite named sources)
- – Hardest for newer/smaller brands to influence
| Engine | Avg Sub-Score | Citation Rate | Best For | Key Signal |
|---|---|---|---|---|
| Perplexity Pro | 44 | 87% | Fresh content publishers | Domain authority + recency |
| ChatGPT (GPT-4o) | 35 | 68% | Comparison content | Review sites + docs |
| Claude 3.5 Sonnet | 31 | 61% | Authority/research content | Cited studies + data |
| Gemini 1.5 Pro | 28 | 30% | Established brands | Wikipedia + KG presence |
Recommendations
Based on our analysis of 527 brands, we identify the following evidence-based recommendations for improving AIS Index scores. These are ordered by typical impact-to-effort ratio, with high-impact, lower-effort actions first.
Publish quantified outcome data
Brands that publish specific numerical outcomes — "customers see 34% reduction in support tickets," "average onboarding time of 2.1 days," "processing 18 million transactions monthly" — score on average 16 points higher on Authority than brands without such data. LLMs cite numbers. Give them numbers to cite.
Create structured comparison and vs. content
Comparison queries ("X vs Y," "alternatives to X") are among the highest-frequency LLM queries in commercial categories. Brands that publish honest, structured comparison pages — including their own weaknesses — are cited more than brands whose comparison content is purely promotional.
Establish presence on third-party authority sources
Wikipedia, G2, Capterra, Trustpilot, and industry-specific review sites all feed LLM training and real-time retrieval. Brands mentioned on three or more such platforms score 22 points higher than those without third-party mentions. Wikipedia presence alone predicts Gemini visibility better than any other single variable.
Use structured data markup (JSON-LD)
Pages with Organization, Product, FAQ, and HowTo schema markup are cited by Perplexity at 1.8x the rate of equivalent pages without markup. ChatGPT also shows sensitivity to structured data in browsing mode. JSON-LD markup is a high-confidence, low-effort signal amplifier.
Publish original research and benchmark data
Original research — surveys, benchmark reports, longitudinal studies — is the highest-value content type for LLM citation. Brands that publish original data are cited at 3.1x the rate of brands publishing only curated or opinion-based content. Even small-scale original research (50-brand benchmark, 500-user survey) drives measurable Authority score improvements.
About This Report
This report was produced by AISearchStackHub using its automated LLM visibility scanning infrastructure. Data was collected from January through April 2026. All 527 brands were analyzed using the same standardized 24-query methodology across ChatGPT (GPT-4o), Claude 3.5 Sonnet, Perplexity Pro, and Gemini 1.5 Pro.
AISearchStackHub provides free AIS Index scans and a Scale plan ($299/mo) that includes the Citation Asset Compounding Engine — an agentic AEO system that generates, tracks, and compounds citeable assets over time. The findings in this report reflect AISearchStackHub's independent research and do not reflect the positions of the LLM providers studied.