Monthly referrals from ChatGPT to B2B sites have soared. Demandbase data shows a sharp increase, but the share of total traffic remains small. The quality and impact of AI-driven visitors are still debated.
Demandbase reports that monthly referrals from ChatGPT to tracked B2B websites jumped from about 645,000 in June 2025 to 2.6 million in June 2026, a 303% increase. This data, based on more than eleven billion visits across 1,584 platform instances, highlights a rapid expansion of a previously minor channel for B2B web traffic.
The growth was not steady. Demandbase notes a sharp spike in May 2026, when referral volume doubled compared to previous months, though the company did not specify the cause. Despite this surge, ChatGPT referrals still represent only a small fraction of total B2B website traffic measured by Demandbase. Other AI assistants showed mixed results: Perplexity saw a decline in referral traffic, while Gemini and Claude remained stable, but Demandbase did not release absolute figures for these platforms.
One reason for the increase may be the nature of B2B buying. While B2C transactions are often fully digital and optimized for self-service, B2B buyers typically require direct contact with vendors for pricing and proposals. This means AI chatbots like ChatGPT can initiate research but rarely complete the transaction, keeping the human element central in B2B sales.
Additional research supports the shift in B2B research habits. According to LinkedIn, 94% of B2B buying groups now use large language models such as ChatGPT or Gemini before engaging with a sales team. This trend moves part of the research process into environments where companies have less control over visibility compared to traditional search or their own websites. However, more ChatGPT referrals do not automatically translate into more business. Studies on AI-driven traffic quality are mixed: one academic analysis of 973 e-commerce sites with $20 billion in combined revenue found that ChatGPT-sourced visitors converted at lower rates than those from traditional channels, while other market reports suggest higher conversion rates for AI-referred users. The impact appears to depend heavily on industry, business model, and measurement approach.
Measuring the true influence of AI referrals remains a challenge. Similarweb data indicates that brands recommended by AI systems are 2.5 times more likely to receive a website visit within a week, but only 8.8% of these visits are directly traceable as referrals. Most occur through subsequent brand searches, meaning standard analytics may undercount AI-driven impact. Monitoring branded search volume can help companies assess the broader effect of AI visibility. For B2B marketers, this requires a shift in measurement strategy: tracking not just direct referrals from ChatGPT, but also brand mentions in AI responses and the resulting search and site activity.
Demandbase cautions that its figures are based on proprietary data and have not been independently verified. The reported 303% growth should be seen as an indicator of channel momentum, not as evidence that ChatGPT is replacing traditional B2B traffic sources. For those interested in how AI platforms are expanding their reach in business contexts, related developments include OpenAI's recent move to target SMBs with new ad formats, as covered in this report on OpenAI's SMB ad platform expansion.
