AI助手成零售新入口，2026年美国AI平台将影响205.7亿美元电商销售

Guilherme Freire Gabriel Reis/Valor A product or service recommendation made during a conversation with an artificial intelligence assistant has become as important to brands as exposure on social media, search engines, or marketplaces. The Black Friday season, which will peak this year on November 27, will test how consumers use AI platforms and whether those interactions actually generate sales. In the United States alone, AI platforms are expected to influence $20.57 billion in e-commerce sales in 2026—nearly four times the $5.4 billion recorded in 2025—according to consulting firm Emarketer. That would represent 1.5% of U.S. online retail sales for the year. Helping brands decipher complex algorithms has become a business opportunity attracting companies including data analytics provider Ibope, software maker Adobe, and e-commerce technology company Dolado. The arrival of AI agents capable of performing tasks such as comparing prices and making purchases is further complicating the landscape. Techniques range from generative engine optimization (GEO) to agentic commerce optimization (ACO). “Instead of beginning their journey on a traditional search engine or browsing a marketplace, a growing share of consumers are asking an AI assistant directly, which responds with lists of stores, products, and selection criteria. That list has become the new gateway to retail. Until now, no one was measuring which brands appeared on it,” said Ibope CEO Antonio Wanderley. The audience-measurement and consumer-data company has launched the Ibope Be.Cited index, which submits real consumer questions to five AI platforms and records which brands appear, their position, and how frequently they are mentioned. Created with Brazilian consulting firm Benext.ai, the new service combines a platform that continuously measures brand visibility, Ibope Be.Cited Intelligence, with strategic recommendations for brands through Ibope Be.Cited Consulting. Wanderley believes the competition is no longer solely about visibility. “For brands, it is no longer enough to know whether they appear in AI responses. They need to understand how frequently they appear, in what position, alongside which competitors, and supported by which sources,” he explained. At Valor’s request, Ibope Be.Cited found that the marketplaces appearing most often in consumer interactions with five AI assistants were Mercado Livre, present in 90% of responses, followed by Amazon at 86%, Shopee at 69%, Kabum! at 51%, and Casas Bahia at 45%. Among the survey’s findings, Wanderley highlighted that “there is no single ‘AI opinion,’” because each large language model (LLM) presents brands differently. “Americanas is the third-most prominent brand on Perplexity, at 57%, but barely appears on Gemini, at 3%,” he noted. Another concern is that brands with different names may lose visibility in AI assistants. “AI treats Magalu, at 42%, and Magazine Luiza, at 41%, as separate brands. Combined, the retailer would be competing for third place.” Consumers already watching for Black Friday promotions may notice that AI “checks before it sells.” According to the Ibope Be.Cited study, 52% of respondents recommend consulting price-comparison sites such as Zoom and Buscapé, while 23% say no retailer is always the least expensive. “From AI’s perspective, the competition is over total cost, not the advertised discount,” the Ibope executive noted. Traffic from generative AI assistants to U.S. retail websites is expected to increase 95% during this year’s Black Friday season compared with last year’s promotional period, Adobe projects. During the 2025 promotional season in the United States, AI platforms drove 693% more consumers to online stores and converted 38% more sales than referrals from other channels. Data on revenue or interaction volume generated by AI was not disclosed. Adobe launched its Brand Visibility software in Brazil in mid-2026. “The idea is precisely to help companies understand how their brands are being mentioned in this new world of agents and LLMs [AI large language models] and influence what LLMs are telling consumers about them,” said Douglas Montalvão, Adobe’s managing director for Latin America. Bradesco is among the platform’s users. In a September interview with Valor, Bradesco Chief Marketing Officer Renato Camargo said the project includes reviewing the bank’s own channels, including its website, which is being adjusted to make information easier for AI to access. Montalvão emphasized that Adobe’s strategy is to demonstrate the financial impact of adopting these solutions. “The success of these initiatives does not lie in purchasing a project or technology, but in adoption,” he said. Digital marketing agency Jellyfish uses AI to simulate different consumer profiles as part of its methodology for measuring brand visibility in LLMs. In fashion and accessories, the category that generated the highest sales volume during last year’s Black Friday season, retailers Shein, Renner, and C&A were mentioned in more than 95% of the responses analyzed across six AI models. Shein and C&A each appeared in 95.77% of responses, while Renner appeared in 97.18%. The agency’s “Share of Model” methodology was based on 71 responses collected weekly between August 17 and September 11, 2026. A total of 213 responses were gathered from questions submitted by synthetic profiles representing Brazilian consumers, 65% of whom were aged 25 to 44. The algorithms offer important clues about the criteria AI models use to highlight particular brands. Product variety accounted for between 28.67% and 43.19% of positive mentions, followed by promotions at 29.14%, trust at 23.94%, and curation at 20.75%, according to Jellyfish’s survey. Conversely, prices, costs, and problems with exchanges and returns were the leading negative factors, with their weight varying by model. “Starting with this year’s Black Friday, brands need to work not only to be found, but also to be considered relevant to consumers’ needs,” said Fabiana Baraldi, Jellyfish’s CEO in Brazil. The spread of artificial intelligence agents capable of purchasing products autonomously with user approval has transformed Brazilian startup Dolado’s strategy for increasing its prospective clients’ sales in marketplaces and e-commerce. Two weeks ago, the e-commerce marketing, logistics, software, and services company partnered with U.S. startup Bodhium Labs, which develops brand-visibility solutions for AI platforms. “With AI agents, the game has changed completely,” Dolado cofounder and CEO Guilherme Freire told Valor. Since September, major technology companies have launched a wave of artificial intelligence agents, including Google’s Gemini Spark, OpenAI’s Dots, Meta Platforms’ Muse, Microsoft’s Copilot Autopilot, Apple’s Siri, and Amazon’s Quick. “These agents do more than recommend products: they can search, compare, negotiate, and make purchases on consumers’ behalf,” Freire said. The entrepreneur added that agents have a broader search scope than artificial intelligence assistants. “This will not be limited to ChatGPT. Personal agents such as Instinct and Muse point to a future in which this experience will be integrated into everyday channels, including messaging applications such as WhatsApp [Meta Platforms’ social network],” he projected. Deodorant brand Mando and credit-card company Visa are among Bodhium Labs’ clients. The startup’s founders include Wharton School AI professor Kartik Hosanagar and former Google DeepMind engineer Krishna Srinivasan. Dolado is selecting companies for the first tests of brand visibility among artificial intelligence agents. They include “one of the country’s largest fashion groups and a major multinational beauty company,” the CEO said, without disclosing their names. Translation: Todd Harkin