More and more people no longer type their purchase questions into Google, but into ChatGPT, Claude or Perplexity. New industry research shows how fast that behaviour is changing, and what it means for brands that depend on being found online.
Commercial intent in AI conversations has nearly doubled
Research from AI answer engine analytics firm Profound shows that the share of ChatGPT conversations with commercial intent rose from 13.9% to 19.2% over the course of a year. Applied to ChatGPT's current conversation volume, that translates to an estimated 28 billion commercial conversations per year.
In other words, nearly 1 in 5 conversations people now have with ChatGPT revolve around looking for a product, service or provider. A year ago that was still fewer than 1 in 7.
This shift matches what we at Trackbase.ai already see in practice. Customers looking for an accountant, a contractor, a beauty clinic or a SaaS solution are increasingly asking an AI model that question instead of a search engine. And the answer that AI gives determines whether your brand shows up at all.
Not every AI model is the same
A second insight is just as important. Different AI models, and even different variants of the same model, draw on different sources. Research into Claude and Claude Code (Anthropic’s developer-focused variant) shows that the two behave, in practice, as fully separate “answer engines,” each citing its own set of pages.
The practical consequence is that a brand visible in ChatGPT is not automatically visible in Claude, Gemini or Perplexity either. And within a single AI provider, the consumer and developer variants can cite different sources again. Brands that take AI visibility seriously need to look, model by model and sometimes variant by variant, at which content and structure actually get picked up.
Regional visibility isn’t a given
It's also worth noting that a strong position in the United States doesn't automatically translate into visibility in Europe. AI models weigh sources and context differently by region, which means European and Dutch brands can't rely on international brand recognition alone. Local, language-specific content and structure remain decisive.
We see this in every implementation project we run: each language version of a website effectively counts as its own point of discovery for AI models, and deserves its own attention.
What does this mean for your brand?
There are three things worth acting on right now. First, take AI search behaviour seriously as a purchase channel. This is no longer a side phenomenon, it's a growing part of how (potential) customers research and decide. Second, optimise per AI model rather than in general. What works for visibility in ChatGPT doesn't automatically work for Claude, Gemini or Perplexity. Third, make sure the technical foundation is solid. Structured data, a well set up robots.txt and llms.txt, and a clear sitemap determine whether AI models can actually read and cite your content in the first place.
At Trackbase.ai we monitor daily how brands score in ChatGPT, Claude, Gemini and Perplexity, and help companies build the technical and structural foundation to become, and stay, visible there.
Curious how your brand is doing in AI search results? Get in touch for a no-obligation analysis.
