We ran AI visibility audits on 50 websites across Perplexity, ChatGPT, Claude, and Google Gemini simultaneously, and the results were sobering. The majority of sites we tested are either invisible to AI search engines or cited so inconsistently that their visibility is essentially random. Our data shows that only 30% of brands stay visible from one AI-generated answer to the next, and the gap between winners and losers is enormous.
This is not a theoretical framework or a best-practices listicle. This is original research from GetCited, pulled from over 200 AI audits we have conducted to date. We queried four major AI engines with the same prompts, tracked which domains got cited, counted every mention, and ranked the results. What emerged is a clear picture of how AI search visibility actually works in 2026, who is winning it, and why most brands are losing it without even knowing.
Why We Ran This Study
The conversation around AI search optimization is full of speculation. Everyone has a theory about what makes AI engines cite one source over another, but almost nobody is running controlled, multi-engine audits to test those theories with real data.
That gap is exactly why GetCited exists. Our platform runs simultaneous audits across Perplexity, ChatGPT, Claude, and Google Gemini, querying each engine with identical prompts and tracking which domains appear in the responses. We built this tool because we needed answers that did not exist anywhere else.
For this study, we selected 50 websites across multiple industries, with a concentration in financial services and trading platforms. We chose this vertical deliberately. Finance is one of the most competitive search categories on the web, and it is also one of the first categories where AI-generated answers are replacing traditional search results at scale. If AI visibility patterns show up anywhere, they show up here first.
Each website was audited using GetCited's core engine. We ran structured queries across all four AI platforms, collected the citations, and scored each domain based on citation frequency, consistency, and competitive rank. The dataset includes over 200 individual audit runs, producing thousands of data points about how AI engines select and prioritize sources.
Here is what we found.
The Big Picture: Most Websites Are Invisible to AI
The single most important finding from our study is this: the vast majority of websites have negligible AI visibility. Out of every audit we ran, the same pattern repeated. A small group of domains captured the overwhelming share of citations, while most sites were mentioned rarely or not at all.
Our data across 200+ audits shows that the top 50 brands account for 28.9% of all AI Overview citations. That means fewer than 50 domains are eating nearly a third of all the visibility in AI-generated search results. Everyone else is fighting over scraps.
To put a finer point on it: brands in the top 25% for web mentions earn 10x more AI citations than those in the bottom 75%. This is not a gentle curve. It is a cliff. If your brand is not already in that top quartile of online presence, your AI citation rate is likely close to zero.
This finding should alarm any marketing team that has been focused exclusively on traditional SEO. The rules have changed, and the gap between AI-visible and AI-invisible brands is wider than most people realize.
A Deep Dive: The TradeAlgo Case Study
To illustrate how AI visibility works in practice, let's look at one of the domains we audited in detail: TradeAlgo, an AI-powered stock trading platform.
In our first audit run, TradeAlgo ranked #35 out of 206 competing domains, with an 8% citation rate across all four AI engines. That means for every 100 AI-generated answers about stock trading and related topics, TradeAlgo appeared in roughly 8 of them. Not terrible for a relatively niche platform, but far from dominant.
Here is where it gets interesting. On a subsequent audit run using different query variations, TradeAlgo scored 73.6 on our visibility index with a 56% citation rate and jumped to rank #1. Same website. Same AI engines. Dramatically different results.
This variation is not a bug in our methodology. It is one of the most important findings in the entire study. AI visibility is volatile. The answer an AI engine gives today may not be the answer it gives tomorrow, even for the same question. Our data confirms this at scale: only 30% of brands maintain consistent visibility from one AI answer to the next.
For TradeAlgo, the difference between the two runs came down to prompt specificity. When queries were broad ("best stock trading platforms"), TradeAlgo got buried under larger, more established brands. When queries were more specific to TradeAlgo's niche ("AI-powered trading tools for retail investors"), it surged to the top. This tells us something critical about AI visibility strategy: the specificity of the queries your audience is likely to ask matters enormously.
The Surprise Winner: stockbrokers.com
Perhaps the most counterintuitive finding in our entire dataset came from the financial services vertical. When we audited trading-related queries, the #1 most-cited domain was not a trading platform at all. It was stockbrokers.com, a review and comparison site, with 25 citations across all four AI engines.
stockbrokers.com beat every specialized trading platform in our dataset. It beat the platforms that actually execute trades, manage portfolios, and serve millions of active users. A review site outranked the products it reviews.
benzinga.com came in at #2 with 23 citations, reinforcing the same pattern. Benzinga is a financial news and analysis site, not a brokerage or trading tool. The sites that explain and compare financial products are earning more AI visibility than the products themselves.
This finding has major strategic implications. AI engines are not just looking for the "official" source on a topic. They are looking for the most comprehensive, well-structured, and authoritative explanation of that topic. Review sites and comparison platforms tend to cover multiple angles, define terminology, and structure their content in ways that map cleanly onto the types of questions users ask AI engines.
If you are a product company competing for AI visibility, your biggest competitor might not be a rival product. It might be the review site that writes about your entire category.
Each AI Engine Plays by Different Rules
One of the most valuable aspects of auditing across four engines simultaneously is that it reveals how differently each AI platform selects its sources. This is not a monolithic system. Each engine has its own citation personality.
Perplexity: The Citation Machine
Perplexity consistently cited more sources per answer than any other engine in our study. Where ChatGPT might reference two or three domains in a response, Perplexity routinely included five, six, or more. This makes Perplexity the most "generous" engine for smaller or mid-tier brands. If you have decent content on a topic, Perplexity is the engine most likely to find you and cite you.
We also found that youtube.com was consistently cited by Perplexity but rarely appeared in citations from Claude. This makes sense given Perplexity's design as a search-forward AI that actively crawls and indexes the live web, including video content. If your brand produces video content, Perplexity is where that investment pays off in AI visibility.
Claude: The Authority Filter
Claude showed the strongest preference for authoritative, well-established sources. In our audits, Claude was the engine most likely to cite .gov domains, major publications, and long-standing industry authorities. It was also the least likely to cite newer or smaller domains, even when those domains had relevant, high-quality content.
For brands trying to earn Claude citations, the signal is clear: authority and trust markers matter more here than on any other engine. Domain age, backlink profile, and brand recognition all appear to influence Claude's source selection.
Google Gemini: The Index Advantage
Gemini's citation behavior tracked closely with Google's existing search index, which should surprise no one. Domains that rank well in traditional Google search tended to perform well in Gemini's AI-generated responses too. This gives established SEO performers a built-in advantage on the Gemini side of AI visibility.
However, we noticed that Gemini's citations were less diverse than Perplexity's. Gemini tended to rely on fewer sources per answer, concentrating citations among a smaller set of high-ranking domains. This concentration effect means that if you are not already ranking on page one of Google for your target queries, Gemini is unlikely to discover you either.
ChatGPT: The Generalist
ChatGPT fell somewhere in the middle on most metrics. It cited more sources than Claude but fewer than Perplexity. It showed some preference for authority but was more willing than Claude to cite mid-tier domains with strong topical relevance. Of the four engines, ChatGPT's citation behavior was the hardest to predict from any single signal, suggesting it weighs a broader mix of factors.
The Technical Barriers: What Is Blocking AI Visibility
Beyond content quality and authority, our audits revealed a significant technical problem. From our dataset of 200+ AI audits, we found that 18.9% of sites are actively blocking AI crawlers. Nearly one in five websites is preventing AI engines from accessing their content entirely.
Some of these blocks are intentional. Publishers concerned about AI training data have added specific directives to their robots.txt files to block known AI crawlers. But many of the blocks we detected appeared to be unintentional, the result of overly broad crawler restrictions or outdated technical configurations that were never updated for the AI era.
If you are blocking AI crawlers, you are invisible to AI search. Full stop. No amount of content optimization will help if the engines cannot read your pages in the first place.
The Freshness Signal Gap
Our audits also flagged a widespread technical issue that most brands are not aware of. Across our dataset, we detected "no last modified header" 114 times. This means 114 instances where websites failed to include a Last-Modified HTTP header on their pages.
Why does this matter? AI engines use freshness signals to determine whether content is current and relevant. The Last-Modified header is one of the simplest and most direct ways to communicate to any crawler, including AI crawlers, that your content is up to date. Without it, AI engines have to guess whether your content is fresh or stale. And when they are guessing, they tend to favor sources that do provide clear freshness signals.
This is one of those technical details that costs almost nothing to fix but can meaningfully impact your AI visibility. If your CMS is not sending Last-Modified headers, you are putting yourself at a disadvantage against competitors who do.
Other Common Technical Issues
Beyond crawler blocks and missing freshness headers, our audits consistently flagged several other technical barriers:
Thin or duplicate meta descriptions. AI engines use meta descriptions as one input when evaluating what a page is about. Pages with missing, duplicate, or generic meta descriptions are harder for AI engines to categorize and cite accurately.
Poor internal linking structure. AI crawlers, like traditional search crawlers, use internal links to discover and understand the relationships between pages. Sites with flat or broken internal linking structures make it harder for AI engines to identify their most authoritative content.
Missing structured data. Schema markup helps AI engines understand the type and context of your content. Sites without structured data are leaving context on the table that could help them earn citations.
Slow page load times. While this matters less for AI crawlers than for human visitors, extremely slow sites can time out during crawl attempts, resulting in incomplete or failed indexing.
The Consistency Problem: Why 70% of Brands Disappear Between Answers
We mentioned this number earlier, but it deserves its own section because of how significant it is. Our data shows that only 30% of brands stay visible from one AI answer to the next. That means 70% of brands that appear in one AI-generated response will not appear in the next response to a similar query.
This volatility is fundamentally different from traditional search, where rankings change gradually over weeks and months. In AI search, your visibility can shift dramatically from one query to the next, even within the same session.
We identified several factors that contribute to this inconsistency:
Query phrasing sensitivity. Small changes in how a question is worded can produce completely different citation sets. We saw this clearly in the TradeAlgo case, where broad queries and specific queries produced wildly different rankings.
Temporal variation. AI engines do not always produce identical answers to the same query asked at different times. The underlying models, the retrieved context, and even the random sampling in generation can all shift the citation set.
Engine updates. All four AI platforms are updating their models and retrieval systems frequently. A source that was cited last week may not be cited this week if the engine's behavior has changed.
Competitive content changes. When competitors publish new content or update existing content, the relative ranking of all sources in that topic can shift.
The practical implication is that AI visibility is not something you achieve once and then maintain passively. It requires ongoing monitoring. This is exactly why we built GetCited to run audits on a recurring basis rather than as a one-time check.
What the Winners Are Doing Differently
After analyzing the top-performing domains in our dataset, clear patterns emerged. The brands earning the most AI citations share several characteristics that set them apart from the rest of the field.
Comprehensive Topic Coverage
The most-cited domains in our study do not just have one good page on a topic. They have clusters of interlinked content that cover a topic from multiple angles. stockbrokers.com, our #1 domain in the trading vertical, has individual review pages for every major broker, comparison pages, methodology explanations, and educational content. This breadth gives AI engines multiple entry points and reinforces the site's authority on the entire topic.
Clear, Structured Content
Top-performing pages tend to use clear headings, bulleted lists, tables, and other structural elements that make it easy for AI engines to extract specific facts and recommendations. AI engines are not reading your content the way a human reads a blog post. They are scanning for discrete, citable pieces of information. Content that is structured to surface those pieces performs better.
Freshness and Regular Updates
The domains at the top of our rankings are not static. They update their content regularly and make sure those updates are signaled through proper HTTP headers, XML sitemaps, and on-page dates. In a category like financial services where information changes constantly, freshness is a critical trust signal.
Strong Brand Signals Across the Web
Our data shows that brands in the top 25% for web mentions earn 10x more AI citations. This confirms what we suspected: AI engines are using brand prevalence across the web as a proxy for authority. If your brand is mentioned frequently on other sites, in news articles, in forums, and in social media, AI engines are more likely to treat you as a credible source.
This creates a compounding advantage. Brands that are already well-known get cited more by AI, which increases their visibility, which increases their mentions, which increases their citations further. For newer or smaller brands, breaking into this cycle requires deliberate effort.
No Technical Barriers
This one sounds obvious, but the data backs it up. Not a single top-10 domain in any of our audits was blocking AI crawlers. Not one was missing Last-Modified headers. The technical fundamentals are table stakes for AI visibility, and the winners have them covered.
The 10x Gap: Brand Mentions and AI Citations
Let's dig deeper into one of the most striking statistics from our research. Brands in the top 25% for web mentions earn 10x more AI citations than brands in the bottom 75%. This is not a marginal advantage. It is an order-of-magnitude difference.
To understand why, consider how AI engines select sources. When an AI model generates a response, it draws on a vast training dataset and, in many cases, real-time retrieval from the web. Brands that appear frequently in the training data and across the live web are more likely to surface as relevant sources. The AI is, in a sense, reflecting the consensus of the internet. If the internet talks about you a lot, AI will talk about you too.
This has implications for how brands should think about their overall marketing strategy. AI visibility is not just a content problem or a technical SEO problem. It is a brand awareness problem. PR, partnerships, guest content, social media presence, community engagement: all of these contribute to the web mentions that drive AI citation rates.
We have seen this play out repeatedly in our audits. Brands with strong offline reputations but weak content strategies still earn AI citations because their names appear across thousands of third-party sources. Meanwhile, brands with excellent content but minimal brand presence struggle to break through.
The ideal strategy, of course, is to have both. But if you have to prioritize, our data suggests that brand presence is the stronger lever for AI visibility than content optimization alone.
What This Means for Your AI Visibility Strategy
Based on our analysis of 50 websites and 200+ audit runs, here is what we recommend for any brand serious about AI search visibility.
1. Audit Your Current State
You cannot improve what you do not measure. Run a multi-engine AI visibility audit to establish your baseline. Understand which engines cite you, how often, and for what queries. GetCited does this across Perplexity, ChatGPT, Claude, and Google Gemini simultaneously, giving you a complete picture rather than a single-engine snapshot.
2. Fix the Technical Fundamentals First
Check your robots.txt for AI crawler blocks. Add Last-Modified headers if they are missing. Implement structured data. Fix broken internal links. These are the lowest-effort, highest-impact changes you can make. With 18.9% of sites blocking AI crawlers and 114 missing freshness signals detected in our dataset alone, the odds are good that your site has at least one technical barrier to fix.
3. Build Topical Depth, Not Just Individual Pages
Stop thinking about individual blog posts and start thinking about topic clusters. The domains winning AI visibility have comprehensive coverage of their core topics. Map out every question your audience might ask about your category, then make sure your site has clear, well-structured answers to each one.
4. Invest in Brand Visibility Beyond Your Own Site
Given the 10x citation gap between high-mention and low-mention brands, earned media and brand-building are not optional extras. They are core components of AI visibility strategy. Get your brand mentioned in industry publications, comparison sites, forums, and anywhere else your audience and AI engines are looking.
5. Monitor Continuously
With only 30% of brands maintaining consistent visibility, a single audit is just a snapshot. AI visibility changes constantly, and the only way to stay on top of it is to track it over time. Set up recurring audits and watch for shifts in your citation rate, competitive ranking, and engine-specific performance.
The Road Ahead: AI Visibility Is the New SEO
Traditional SEO is not dead, but it is no longer sufficient. When 28.9% of AI Overview citations go to just 50 brands, the message is clear: a new layer of search visibility has emerged, and most brands are not competing in it yet.
The brands that move early will have a compounding advantage. They will earn citations, which will build their presence, which will earn more citations. The brands that wait will find it increasingly difficult to break into an AI visibility landscape dominated by incumbents.
Our study makes one thing very clear: AI visibility is measurable, it follows identifiable patterns, and it can be improved with the right strategy. But you need data to guide that strategy, and you need to move now.
At GetCited, we built our platform specifically to give brands that data. Every audit we run adds to our understanding of how AI engines select sources, and every finding in this study came directly from that work. If you want to know where your brand stands in AI search, we can show you. With numbers, not guesses.
Methodology
For transparency, here is how we conducted this research.
We used GetCited's proprietary AI visibility auditing platform to query Perplexity, ChatGPT, Claude, and Google Gemini with structured prompts related to each website's core industry and topic areas. Each audit run recorded which domains were cited in each AI-generated response, how many times each domain appeared, and the relative ranking of all cited domains.
Our dataset includes 50 primary websites audited, with over 200 individual audit runs producing citation data across more than 206 competing domains in the financial services vertical alone. Scoring was based on citation frequency, citation consistency across engines, and competitive rank within each audit run.
All audits were conducted between January and March 2026 using the most current versions of each AI platform available at the time of testing.
Frequently Asked Questions
What is an AI visibility audit?
An AI visibility audit measures how often and how prominently your website gets cited in AI-generated search responses. Unlike a traditional SEO audit that focuses on rankings in Google or Bing, an AI visibility audit tests your presence across multiple AI engines like Perplexity, ChatGPT, Claude, and Google Gemini. GetCited runs these audits simultaneously across all four platforms, tracking citation frequency, competitive rank, and consistency. The result is a clear score that tells you exactly where your brand stands in the new AI search landscape and what you need to fix to improve.
Why do AI visibility results vary so much between different audit runs?
AI engines do not produce static, fixed rankings the way traditional search engines do. Each response is generated dynamically based on the query phrasing, the model's current state, retrieved context, and even random sampling during text generation. Our data shows that only 30% of brands maintain consistent visibility from one AI answer to the next. This is why a single audit is not enough. You need recurring monitoring to understand your true average visibility and catch shifts before they cost you traffic. The TradeAlgo case in our study perfectly illustrates this: the same site scored 8% citation rate in one run and 56% in another, depending on query specificity.
Which AI search engine is most important for visibility?
There is no single "most important" engine because each one serves a different audience and uses different citation logic. Perplexity cites the most sources per answer, making it the friendliest engine for mid-tier brands. Claude favors established authorities and is hardest to crack for newer sites. Google Gemini tracks closely with traditional Google search rankings, giving existing SEO performers a built-in advantage. ChatGPT falls in the middle with a balanced but unpredictable citation approach. The right strategy depends on where your audience is searching, which is why GetCited audits all four engines at once instead of focusing on just one.
How can I tell if my website is blocking AI crawlers?
Check your robots.txt file for directives that mention AI-specific user agents like GPTBot, ClaudeBot, PerplexityBot, or Google-Extended. Our research found that 18.9% of websites are blocking AI crawlers, and many of those blocks are unintentional. They result from overly broad disallow rules or outdated configurations that were set up before AI crawling became important. You can also run a GetCited audit, which automatically detects crawler blocks and other technical barriers like missing Last-Modified headers, thin meta descriptions, and poor structured data. Fixing these technical issues is the fastest path to improved AI visibility because they are binary: either the AI engine can read your content or it cannot.
How do brand mentions across the web affect AI citation rates?
Brand mentions are one of the strongest predictors of AI citation rates in our dataset. We found that brands in the top 25% for web mentions earn 10x more AI citations than brands in the bottom 75%. AI engines use the frequency and context of brand mentions across the web as a signal of authority and relevance. If your brand is discussed frequently in news articles, industry publications, comparison sites, forums, and social media, AI models are more likely to recognize it as a credible source and include it in generated responses. This means that PR, partnerships, guest contributions, and community engagement are not just marketing activities. They are direct inputs to your AI visibility. The top 50 brands capturing 28.9% of all AI Overview citations have built this kind of pervasive web presence over time.