ChatGPT is developing into a relevant access point for B2B offerings. An analysis by Demandbase shows a 303 percent increase in redirects to company websites within a year. At the same time, its share of total web traffic remains small, and the data on visitor quality is inconsistent.
According to Demandbase, the number of monthly redirects from ChatGPT to the B2B websites tracked by the company increased from approximately 645,000 in June 2025 to 2.6 million in June 2026. This represents a rise of 303 percent. The analysis by Labs by Demandbase encompasses more than eleven billion website visits across 1,584 platform instances. However, it is noteworthy that the increase was not uniform: In May 2026, the volume doubled compared to the preceding months. Demandbase does not explain why this jump occurred.
This means a previously small traffic channel is growing rapidly. The 2.6 million ChatGPT referrals still represent only a small fraction of the total volume tracked by Demandbase. The performance of other AI assistants also varies: according to Demandbase, Perplexity lost referral traffic, while Gemini and Claude remained largely stable. However, Demandbase does not publish the absolute figures for these providers.
The growth may seem surprising at first glance, but it has an inherent logic. While many B2C inquiries, including product purchases, are fully digital and optimized for self-service, B2B customer inquiries typically require direct contact with the supplier to obtain price quotes. Therefore, the AI chatbot cannot fully fulfill the task at hand.
AI begins B2B research
Other analyses also show that AI systems are being used even before direct contact with a supplier. A recent LinkedIn study found that 94 percent of B2B buying groups use large language models like ChatGPT or Gemini before speaking with a sales team. This shifts part of the research process to an environment where companies no longer directly control their visibility through traditional search results or their own website.
However, it’s incorrect to conclude that more ChatGPT referrals automatically translate to more business. Studies on the quality of AI-generated traffic have yielded mixed results. One academic study of 973 e-commerce websites with a combined revenue of $20 billion found that ChatGPT-mediated traffic lagged behind several traditional channels in terms of conversion rate and revenue per session. At the same time, other market analyses report higher conversion rates for AI-generated visitors. The results, therefore, appear to depend heavily on the industry, business model, and measurement method.
Not every AI effect will be visible as a referral.
Another problem lies in measurement. According to Similarweb, brands recommended by AI systems were 2.5 times more likely to be visited within seven days. However, only 8.8 percent of these AI-influenced visits came directly from a recognizable referral. A large proportion subsequently occurred via brand searches. Traditional web analytics can therefore underestimate the influence of ChatGPT and other assistants. Conversely, monitoring brand searches can help to evaluate the impact of AI presence.
For B2B marketing, this means a shift in measurement logic. The crucial factor is no longer solely how many visitors ChatGPT sends to a website. Companies must also monitor whether their brand appears in AI responses and whether this visibility subsequently translates into search queries, website visits, and business leads. Demandbase itself, however, points out that its figures are derived from a proprietary dataset and have not been independently verified. The 303 percent figure is therefore primarily an indicator of the channel’s dynamism – not yet proof that ChatGPT is replacing traditional B2B traffic.
