When AI Agents Do the Shopping
AI assistants are transforming retail, acting as high-intent intermediaries that pick products long before a human reaches the aisle.
AI assistants are transforming retail, acting as high-intent intermediaries that pick products long before a human reaches the aisle.

For generations of brand builders, retail strategy meant winning the human eye: commanding prime shelf placement, crafting persuasive packaging, and bidding for visibility on search engines. Today, that familiar dynamic is fragmenting. Consumers are increasingly deputizing artificial intelligence agents to browse, evaluate, and purchase goods on their behalf, introducing an opaque algorithmic intermediary between product and consumer.
The rules governing this new digital shelf are inconsistent across platforms. According to Jellyfish’s Share of Model tool, which tracked recommendations across eight product categories in the United States, Britain, Australia, and Singapore, visibility hinges heavily on which assistant a consumer opens. When prompted for toy recommendations, Alexa for Shopping drew products from 177 brands, yet virtually every link originated exclusively from Amazon’s own store. ChatGPT distributed identical inquiries across 24 distinct retailers, while Google’s AI Mode cited 37. Being stocked across diverse retailers matters little if a shopper’s chosen assistant refuses to look beyond its own walls.
Channel distribution also shifts based on technical friction behind the scenes. Agency Tinuiti found that Microsoft’s Copilot cited major retailers in 34% of its shopping prompts, triple the rate seen on Google’s AI Mode and eight times that of ChatGPT. Part of this retailer preference stems from platforms like Reddit blocking Copilot’s web crawlers, creating empty real estate that commercial storefronts fill. Yet across all assistants tracked, retail citations still slipped to 14% by July.
Marketers are moving quickly to adapt. Dentsu Creative’s survey of 1,950 global marketing leaders revealed that 59% are optimizing content and SEO for AI-driven engines, tying with social commerce and personalized messaging as their top innovation priority. Half are investing in agentic commerce to support customer acquisition.
The urgency is practical. Adobe tracked over one trillion retail site visits and found AI-referred shoppers generated 53% more revenue per visit than standard traffic, reversing a trend from last year when non-AI visitors were worth significantly more. These buyers arrive pre-qualified, having used conversational prompts to narrow decisions before clicking.
This shift will likely accelerate as social networks integrate shopping agents. Cyber security firm HUMAN Security recently reported that Meta’s Muse logged 5.6 million requests in under two days, a volume ChatGPT took 11 months to achieve. Muse now generates 72% of all daily agent requests tracked by the firm, with 84% aimed at product and search pages and a checkout conversion rate 25% higher than competing bots, powered by distribution across Instagram, WhatsApp, and Facebook.
When an autonomous agent filters choices before a shopper ever sees an ad, conventional brand loyalty is tested. The challenge facing modern marketers is no longer just persuading people, but teaching machines why a brand deserves to be shortlisted in the first place.