The new sales funnel: from search engines to AI agents
The e-commerce ecosystem is experiencing an accelerated shift in the origin of its audiences. According to recent data from the Tiendanube Mexico platform, traffic driven by artificial intelligence models like ChatGPT, Gemini, and Claude towards online stores recorded a year-on-year growth of 396%. This increase, which multiplies the previous volume of visits almost fivefold, suggests that conversations with intelligent agents are displacing traditional organic search as the primary source of product discovery. David Soto, product marketing manager at Tiendanube Mexico, explains that this channel has become one of the fastest-growing for the sector, directly competing with the territory that previously belonged exclusively to Google or social media advertising.
The differential between algorithmic curiosity and final conversion
Despite the explosion in visibility, reaching a store's interface does not guarantee a transaction. Metrics reveal that of every 100 visits driven by an artificial intelligence recommendation, only between three and five conclude in an effective sale. This conversion rate of 3% to 5% underscores that, at the current stage of technology, the final decision-making and payment process remain strictly under the control of human judgment. This phenomenon of massive growth in traffic with moderate conversion is not exclusive to the local market; Adobe Analytics documented a 393% increase in AI-driven traffic to retail sites in the United States during early 2026.
Operational efficiency and the race for autonomous checkout
As consumers explore the use of assistants to compare prices or receive suggestions, businesses have accelerated their own technological adoption to optimize the offering. In Latin America, about 95,000 of the 180,000 stores operating under Tiendanube are already actively employing integrated artificial intelligence functions. These tools have allowed saving approximately 900 million hours of work on catalog updates and search engine positioning tasks. In terms of revenue recovery, assistants like Chat Nube have managed to raise sales recovered from abandoned carts by 30% through interactions on WhatsApp and Instagram.
The industry anticipates that the next evolutionary leap will occur when the buying cycle is completed entirely within the AI agent. David Soto estimates that by 2027 the first fully autonomous checkout systems could be available, where the assistant executes the transaction without the need for a final redirect to the physical or digital store. In light of this scenario, the Mexican Online Sales Association warns that companies must prepare to compete not only for people's attention but also for the preference of AI algorithms, which prioritize structured information, verifiable reputation, and price consistency.
From the editorial perspective of next+, the exponential growth of AI-assisted traffic represents a structural change in product discovery that forces brands to optimize their presence to be "readable" by machines. What we are observing is a transitional stage where the operational efficiency of businesses has matured faster than users' purchasing autonomy. For executives in Mexico and Latin America, the urgency lies not only in attracting this massive traffic but also in preparing the necessary data infrastructure for the moment when AI agents become the final decision-makers in corporate and personal transactions.