ChatGPT has yet to become a major force in online shopping despite widespread predictions that artificial intelligence would transform the way consumers buy products, according to new research from Frankfurt School of Finance & Management.
The study analysed traffic from ChatGPT to 973 e-commerce websites with combined revenues exceeding $20 billion between August 2024 and July 2025. Researchers found that referrals from large language models accounted for less than 0.2 per cent of total website traffic, highlighting the continued dominance of established marketing channels such as search engines, email marketing and affiliate networks.
Researchers compared ChatGPT referrals with traditional sources including organic search, paid search, paid social media and email campaigns. While ChatGPT generated stronger conversion rates than paid social advertising, it still lagged behind every other major marketing channel in both conversion performance and revenue generated per session.
The findings suggest that generative AI is unlikely to replace conventional customer acquisition methods in the near future.
However, the research did identify one area where ChatGPT is already proving valuable: helping consumers make complex purchasing decisions. Products that require extensive research, comparisons or expert guidance appear to benefit from AI-assisted recommendations, while simpler purchases show little advantage.
Professor Christian Schulze, Professor of Marketing at Frankfurt School, said the results challenge assumptions that ChatGPT has already reshaped e-commerce.
“Our findings contradict the idea that ChatGPT has already transformed online shopping — but the real story is in the product categories,” he said. “The study shows where ChatGPT already has substance in e-commerce and where it remains largely hype.”
Schulze said consumers buying more complex products can benefit from ChatGPT’s ability to organise information, compare alternatives and simplify decision-making. For routine purchases, however, the technology currently offers limited additional value.
The study also points to opportunities for AI platform developers. Researchers argue that improvements in product availability, recommendation quality and checkout processes will be essential if large language models are to become more influential shopping tools.
Several AI companies have already begun introducing features such as instant checkout and autonomous shopping assistants since the study concluded, signalling efforts to strengthen the role of generative AI in e-commerce.
For retailers, the researchers recommend viewing ChatGPT as a complementary marketing channel rather than a replacement for existing strategies, particularly for businesses selling products that require customers to conduct significant research before purchasing.
The study, titled ChatGPT Referrals to E-Commerce Websites: How Do LLMs Compare Against Traditional Channels?, has been published in the journal Marketing Science.
