If your brand launched after 2023, most AI models don't know you exist
Five of the 10 major AI models can see you through web search. The other five run on training data and don't know newer brands exist. That's a structural divide.
If your brand launched after 2023, most...
TL;DR
Across three reports and 91 brands, we found a structural divide in AI visibility no amount of content optimization can fully overcome. Brands with significant web presence before 2024 show up on all 10 AI models. Brands that launched or gained traction after that point are nearly invisible on training-only models and only appear on the models that search the web in real time. The AI search landscape isn't one leaderboard. It's two.
The two-tier AI search landscape
AI models fall into two categories.
Web-search models (ChatGPT with search, Perplexity, Claude, Google AI Overview, Google AI Mode) can access current web content when generating responses. They see your latest blog posts, documentation, community discussions, and comparison articles. If your brand is being discussed online right now, these models can find you.
Training-only models (Gemini, DeepSeek, Mistral, Qwen, and base versions of other models) generate responses based on patterns learned during training. Their knowledge has a cutoff date. If your brand wasn't well-represented in web content before that cutoff, these models won't recommend you.
The data across three categories
In outbound sales, Clay scores 21.1 overall but gets nearly all of that from web-search models. On training-only models like Mistral and Qwen, Clay barely registers. The same holds for Instantly (19.0) and Smartlead (16.3). These are brands with significant real-world adoption that are invisible to half the AI landscape.
In coding tools, the effect is more dramatic. Claude Code (released February 2025) scores 33.7 overall but only appears on 8 of 10 models. Codex (released April 2025) scores 16.4 and appears on 9 of 10. On the training-only models that do mention them, their scores are substantially lower than on web-search models.
Training-only models account for 35 to 65 percent of a brand's total visibility share depending on the category. If you're invisible on these models, you're missing a large chunk of the AI recommendation surface.
This is structural, not fixable by content alone
You cannot retroactively insert your brand into a model's training data. No amount of blog posts, documentation, or comparison articles will make Gemini's base model aware of a product that launched in 2025. That awareness requires the model to be retrained with fresher data.
What you can do is maximize your visibility on web-search-enabled models. That means building content that's comprehensive, well-structured, and updated frequently. It means appearing in the sources web-search models trust: vendor-neutral comparisons, community discussions, technical documentation.
For newer brands, the strategic calculation is straightforward. Five of the 10 major AI models can see you through web search. The other five cannot. Optimize aggressively for the five that can, and wait for the next training cycle on the five that cannot.
The window matters
AI models get retrained. The next generation of Gemini, DeepSeek, and Mistral will include content from 2025 and 2026. Brands building a strong web presence now will be included in that training data. Brands that aren't will face the same invisibility problem on the next generation of models.
The training data cutoff is not a permanent disadvantage. But it's a time-sensitive one. The content you publish today shapes whether the next generation of AI models knows you exist. The first move is finding out which models can already see you.
Observability report · Outbound Sales report · AI Coding Tools reportFrequently asked questions
Why can't some AI models see my newer brand?
Training-only models (Gemini, DeepSeek, Mistral, Qwen) generate answers from data with a cutoff date. If your brand wasn't well-represented in web content before that cutoff, those models have no knowledge of you and won't recommend you.
Which AI models can see recent brands?
Web-search-enabled models: ChatGPT with search, Perplexity, Claude, Google AI Overview, and Google AI Mode. They pull current web content, so they can find brands being discussed online right now.
Can I get into a model's training data after launch?
Not retroactively. But you can build the web presence that gets you included in the next training cycle, and maximize visibility on web-search models in the meantime. The cutoff is time-sensitive, not permanent.
How much visibility am I losing on training-only models?
Training-only models account for 35-65% of a brand's total visibility share depending on category. If you're invisible there, you're missing a large slice of the AI recommendation surface.
Renown is an AI visibility platform that tracks how AI models talk about your brand across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews.
Related Articles
AI visibility is the new SEO. Here's...
AI visibility is the new SEO. Here's the data to prove it
In 2010, if you weren't on page one of Google you were invisible. AI search is at that same inflection point, and the data from 91 brands proves it.
How we measure AI visibility
How we measure AI visibility: querying 10 LLMs with 100 prompts per category
10 models. 100 prompts per category. Scored for mentions, recommendations, citations, and sentiment. Here's exactly how the data works, including what it can't tell you.
What AI trusts
What AI trusts: a citation analysis across 3,000 AI responses
AI search engines cite sources. We tracked which ones across 3,000 responses. The hierarchy is clear, and most marketing content sits at the bottom of it.
