Benchmark: Why 60% of Major Brands Are Invisible on ChatGPT and Perplexity
The Great Marketing Denial
There's a scene that plays out every Monday morning in marketing offices around the world. The SEO manager opens their dashboard, sees the brand ranking number one for twenty strategic keywords, and smiles. A good week ahead. What they don't look at — because the tool doesn't exist in their stack yet — is the query their next customer just typed into ChatGPT: 'What's the best solution for managing my product catalog online?' And in the perfectly written 400-word response from the AI assistant, their company doesn't appear anywhere. Three competitors do. Meanwhile, the dashboard still shows 'Position 1' in bright green.
This gap between perceived and actual digital reality is at the heart of a study we conducted across 500 representative companies — 250 e-commerce players and 250 SaaS vendors — over a six-month period. The goal: measure the divergence between traditional SEO visibility and what we now call 'synthetic visibility' — a brand's ability to appear in responses generated by conversational AI engines. Spoiler: the gap is staggering. And it's going to keep widening.
5,000 Queries and Gallons of Coffee: How We Ran the Study
Running a serious AI visibility benchmark is no walk in the park. Our methodology involved submitting over 5,000 transactional and comparative queries to the major conversational AI assistants: ChatGPT (GPT-4o), Perplexity AI, and their mobile variants. Every query was phrased the way a real user would: 'What's the best e-commerce platform for small businesses?', 'Compare CRM tools for a 20-person sales team', 'Which payment solutions do you recommend for a B2B website selling across Europe?' Nothing robotic, everything natural — or as close as we could get.
For each company in the panel, we tracked citation frequency, position in the response (top recommendation versus secondary mention), associated sentiment (positive, neutral, negative), and the quality of the AI's description. We also cross-referenced these results with each brand's Google rankings on equivalent keywords to quantify the gap between SEO performance and GEO performance. The result is an unprecedented snapshot of brand presence in the generative ecosystem — a ranking that often bears no relationship to their Google position.
60%, 12%, 15%: Three Numbers That Redraw the Digital Marketing Map
Let's start with the bluntest stat: 60% of the companies in our sample — brands that have invested years of work and significant budgets to dominate Google's first page — simply never appear in AI-generated responses. Not once across 5,000 queries. As far as generative engines are concerned, they don't exist. It's the equivalent of building a stunning flagship store in the heart of the city, with perfect window dressing and a trained team — only to find that everyone's GPS keeps sending customers to the competitor next door.
The second number may be even more unsettling for SEO teams: only 12% of sites that hold Top 3 positions on Google for strategic keywords are spontaneously cited by AI assistants on equivalent queries. In other words, dominating traditional search engines provides virtually no guarantee of existing in generative responses. And finally, only 15% of studied brands are regularly recommended as a top pick by AI. These companies don't just exist in responses — they govern them. They're the ones capturing the leads, the trust, and the conversions at the exact moment a decision crystallizes in the user's mind.
The E-commerce / SaaS Paradox: Why Software Vendors Do Slightly Better
Our study reveals a surprising divide between the two sectors analyzed. On the SaaS side, 42% of vendors show minimal visibility in AI responses — a mediocre score in absolute terms, but a relative win in this bleak landscape. The explanation is fairly straightforward: B2B software naturally accumulates documentary traces on high-authority third-party platforms — G2, Capterra, Trustpilot, ProductHunt, industry newsletters, specialist blogs. These sources are exactly the type of 'external validation' that LLMs use to justify a recommendation. A SaaS solution well-rated on G2 with 300 detailed reviews has infinitely better odds of being cited by ChatGPT than a brilliantly optimized corporate website that exists in a vacuum.
E-commerce, by contrast, faces a deeper structural problem. 68% of the B2C brands in our panel are absent from direct purchase recommendations on ChatGPT once the query goes beyond pure brand search. When a user asks 'which online store has the best hiking boots for women,' AI responses lean heavily on editorial guides, press comparisons, and community forums — not the brands' own product pages. A retailer that exists only on its own domain, with no external editorial footprint, is simply transparent to generative models. And transparent, in this context, doesn't mean modern — it means nonexistent.
Anatomy of Invisibility: Three Traps Most Brands Fall Into
The first — and by far most widespread — factor behind AI invisibility lies in the architecture of digital content itself. Landing pages and product pages optimized for Google follow an attraction logic: they're designed to capture keywords, drive traffic, and convert. They are tools of algorithmic seduction. But an LLM works differently: it seeks to understand, synthesize, and answer precisely. It values logical structures (comparison tables, feature lists, measurable figures), explicit answers to real user questions, and content that takes a clear position. A purely promotional piece — loaded with superlatives and light on factual data — will be systematically bypassed in favor of a more neutral, more structured third-party source, even if that source ranks lower on Google. The irony is brutal: the more you've optimized to please Google, the greater the risk of being invisible to AI.
The second trap is the 'proprietary bubble' — betting everything on your own website while neglecting the surrounding ecosystem. To make confident recommendations, ChatGPT and Perplexity cross-reference official site information with data spread across the entire web: forums, trade press, authority directories, independent reviews, knowledge bases. A brand that has mastered its site but ignored its presence in industry media loses the trust signal that LLMs use to validate suggestions. AI doesn't recommend what it sees on a single website — it recommends what achieves consensus across its training universe. Third trap: assuming ChatGPT and Perplexity operate identically. Perplexity prizes very recent sources and highly structured content with hard numbers. ChatGPT relies more on global brand notoriety and its recurrence in sectoral ecosystems. A brand can be perfectly visible on one and completely absent from the other — making cross-platform auditing essential.
The Window Is Closing: Why Acting Now Beats Waiting Until 2028
The history of the web has an annoying habit of repeating itself. In 1998, a handful of visionaries understood that search engine optimization would become a major strategic issue. Their competitors looked on with condescension while investing in Yellow Pages. In 2012, the first brands to structure their social presence built audiences that are impossible to replicate today at any price. In 2026, we're standing at exactly the same crossroads. Brands investing in GEO (Generative Engine Optimization) today are building a presence that competitors will take years to catch up to — and as always in these windows, the majority will wait until they have irrefutable proof it's necessary, by which point any competitive advantage will be long gone.
But before acting, you need to measure. No traditional tool tells you whether your brand appears in AI responses, in what context, with what sentiment, alongside which competitors. Google Analytics tells you how many visitors arrive on your site. It doesn't tell you that your brand wasn't mentioned in 300 conversations where your next customer was looking for exactly what you sell. That's precisely the gap my-ai-visibility.com is built to fill — a platform dedicated to auditing and monitoring brand visibility within generative engines. Measuring your AI share of voice, identifying misinformation the AI may be spreading about your brand, mapping the external sources that shape its perception: these indicators have become as critical as your Google click-through rate. While 60% of the market is still flying blind, the brands that act now are building the most underestimated — and most durable — competitive advantage of the next decade.