ChatGPT Drives Less Than 1% of Google Traffic Insights

According to a study by Marie Haynes, ChatGPT brought websites only 0.62% as many visits as Google's unpaid search results. The study looked at Google Analytics 4 (GA4) data from 100 websites over 12 months, covering more than 145 million visits.

When visits from other large language models (LLMs), which are AI tools like ChatGPT, were included, the number rose slightly to 0.73% of Google's unpaid search traffic.

These percentages compare AI traffic only with Google's unpaid search traffic, not with all visits from every source.

Haynes called the current focus on generative engine optimization (GEO), which aims to help businesses appear in AI search results, a “distraction.”

The traffic numbers support her point that AI tools currently send far fewer visitors than Google. However, they do not answer the bigger question of how important being visible in AI search results might become in the future.

For now, Google remains the main source of website traffic. Marketers need to keep this in mind while also watching how AI search could change the way people find brands and services.

ChatGPT Visitors Are More Likely to Take Action, but There Are Far Fewer of Them

Among the 55 websites that tracked conversions, 13.8% of visitors coming from ChatGPT completed a desired action, such as making a purchase or signing up. In comparison, only 3.02% of visitors from Google's unpaid search results completed such an action.

Although ChatGPT visitors were more likely to convert, there were not enough of them to make up for the large difference in traffic. Google's unpaid search results still produced 27 times more conversions than all the AI language models combined.

“The AI search that is actually directing customers to businesses today isn’t ChatGPT. It’s Google,” Haynes said in her findings.

Growth in Website Visits From ChatGPT Has Slowed

The study found that visits coming from ChatGPT to the websites being tracked stopped growing at a very fast rate by August 2025. Haynes reported that these visits began to fall after that.

Haynes studied websites of different sizes. However, the study did not give the exact dates covered by the data, identify the types of businesses involved, or explain exactly what counted as a conversion. It also did not look at how Google's unpaid search traffic differed across the websites.

Source: Martech.org.