How AI Agents Are Changing Online Shopping
As AI agents compare and buy on our behalf, businesses face a new question: Who carries the risk when the customer is code?
The rise of AI agents is transforming the landscape of online shopping, with companies like Amazon and Perplexity clashing over the implications. Meta's Muse AI agent, which debuted in September 2025, can perform tasks like sending emails, filling out forms, and booking travel, even making purchases with Stripe's Link wallet. The basic version is free and quickly became a hit on the U.S. free-app charts.
In response, Amazon banned Muse agents from its platform, arguing that outside applications should respect retailers' decisions about participation.
This isn't Amazon's first confrontation with AI shoppers. Earlier in 2024, Perplexity's AI-powered search engine launched Buy with Pro, enabling subscribers to purchase recommended products with a single click. Perplexity's chief business officer, Dmitry Shevelenko, defended the tool, stating it acts as an agent on behalf of the user, complementing Amazon's role as a merchant.
However, the conflict escalated in 2025 when Perplexity introduced Comet, an AI-powered browser, leading Amazon to introduce a technical block. Comet developers quickly circumvented this, prompting Amazon to issue a cease-and-desist order, which eventually led to a lawsuit.
The issue goes beyond just banning agents; it raises questions about accountability in commercial decisions. Advertising revenue is at stake, as traffic generated by agents complicates how Amazon measures advertiser payments. Advertisers pay for human attention, and Amazon's advertising business generated $19.8 billion in the second quarter of 2026.
A preliminary injunction in March 2026 favored Amazon, but a federal appeals court vacated it in August, ruling that Perplexity's customers, not the agents, were making the purchases.
This conflict underscores a broader issue: when the entity making a commercial decision cannot be held responsible, what happens? Every purchase involves information and assurances, but AI agents lack this accountability. Forrester's surveys reveal that 94% of business buyers have used generative AI during their most recent purchase, up from 89% the previous year.
While 67% prefer making purchases without sales representatives, 69% seek human input to verify AI's recommendations. This suggests a division of labor between machines providing information and humans ensuring accuracy and risk management.
Forrester's findings indicate that 13 people—inside and outside a company—now influence average business purchases. AI discovers and assesses potential, while humans approve decisions and intervene when needed. Amazon can counteract AI agents by blocking them and developing its own. However, this approach is limited to companies with the resources to build such capabilities.
The deeper issue lies in talent—new roles, incentives, and performance metrics will be necessary as machines take over the discovery and assessment aspects of sales, leaving humans to focus on human interactions and relationship-building.
Written by urgent.news from Newsweek's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
