New research from Tidio reveals a critical measurement gap in e-commerce: while artificial intelligence influences approximately half of consumer purchase decisions, standard attribution models credit AI with less than 1% of web traffic. This discrepancy, detailed in the report AI in E-Commerce in 2026: The New Shopping Funnel, suggests brands are making budget decisions based on incomplete data, potentially missing a major driver of consumer demand.
The report, drawing on over 60 sources including McKinsey, Contentsquare, and Similarweb, identifies a mechanism it terms "dark AI." A consumer asks an AI assistant for product recommendations, receives a shortlist, and then navigates directly to a recommended brand via a new browser tab or branded search. The resulting web session is typically registered as direct or organic traffic, leaving the AI that initiated the journey with no attribution. This creates influence that is commercially real but analytically invisible to current marketing models.
The financial stakes of this gap are substantial. McKinsey projects $750 billion in U.S. revenue will flow through AI-powered search by 2028. Brands that fail to prepare for this shift risk losing 20 to 50 percent of their traditional search traffic. Further analysis by Morgan Stanley estimates AI agents will influence between $190 billion and $385 billion in U.S. e-commerce spending by 2030.
Data on the limited sessions that are tagged as AI-referred suggests the undercounting is significant. Similarweb's analysis of U.S. retail data finds ChatGPT-referred sessions convert at 11.4%, the highest rate of any measured channel. This surpasses direct traffic at 10.2%, paid search at 9.3%, and organic search at 5.3%. Such a conversion premium implies that tagged AI referrals represent only a high-intent fraction of a much larger pool of AI-influenced consumer journeys.
The attribution gap appears to be widening. TollBit's analysis of AI bot behavior across publisher sites found click-through rates from AI applications dropped nearly threefold over 2025, from 0.8% in the second quarter to 0.27% by year-end. This indicates AI platforms are consuming more content while generating proportionally fewer outbound clicks, further obscuring their influence in standard analytics.
"Brands making budget decisions based on last-click attribution are optimizing for a measurement system that cannot see what is actually driving demand," said Tytus Gołas, Founder and CEO of Tidio. He noted that the inputs determining AI visibility—such as feed completeness, structured data, and review coverage—often live across multiple organizational teams with no clear ownership, because the return on investment remains invisible.
Concurrently, protocol infrastructure is being built to formalize AI's role in transactions. Initiatives like Google's Universal Commerce Protocol, OpenAI's Agentic Commerce Protocol, and Visa's Trusted Agent Protocol are creating standardized rails for AI agents to complete purchases on behalf of consumers. Consumer readiness for this shift is building rapidly. Omnisend's longitudinal research found reluctance to allow AI to complete transactions dropped from 66% to 32% in just five months between February and July 2025.
The full research report, AI in E-Commerce in 2026: The New Shopping Funnel, is available for download at https://www.getlyro.ai/reports/ai-in-ecommerce.
