According to the PYMNTS Intelligence report "The AI Power User Gets Pickier," 23% of consumers in the US have used AI to search for products in the past year. Among those using AI for research, that number is 93%. Researching and comparing products is the only one of ten measured shopping activities that showed a noticeable increase. AI recommendations may soon become as important as search rankings.
Discussions about AI in commerce often focus on agents that buy and pay on their own. But real changes happen earlier – at the selection stage. Consumers don’t need AI to click "Buy." The value lies elsewhere: it helps navigate decisions before purchase – which laptop has better battery life and power, whether the premium version is worth an extra $300, which sneakers are suitable for knee issues. Regular search provides information. Generative AI synthesizes it according to human limitations. Before buying something for the consumer, AI decides who gets the chance to sell.
For brands, the question shifts from "how high do we rank?" to "are we mentioned at all?" Recommendation interfaces compress several stages of the funnel into one interaction: discovery, familiarity, and comparison happen in one conversation, and the consumer arrives at the retailer with a short list. By the time of the click, the competition is already behind. This resembles a shift to SEO, but the goal is different: SEO asks if a page can be found, while AI discovery asks if a product can be understood well enough to compare and recommend. AI does not replace search engines and marketplaces, but control over commerce goes to those who control the gateway. Dominance in AI commerce does not necessarily require owning the transaction: by determining which products make it to the list, they shape what consumers search for and where purchase intent lies.