Amazon AI agent scraped merchant catalogues without consent in "Buy for Me" trial
Amazon's Buy for Me AI agent scraped product catalogues from hundreds of independent online merchants without consent and made them purchasable through the Amazon app, generating incorrect descriptions, wrong images, and exposing wholesale pricing. Merchants reported high chargebacks, unauthorised dropshipping arrangements, and loss of control over their brand identity and return policies.
What happened
Amazon expanded trials of Buy for Me, an agentic AI tool integrated into its Rufus assistant, allowing customers to find products on external websites hosted on Shopify, WooCommerce, or Squarespace and complete purchases within the Amazon app.
Hundreds of independent merchants discovered their entire product catalogues had been scraped and listed on Amazon without their knowledge or permission. The AI generated incorrect product descriptions, used unauthorised images, and sold items reported as out of stock.
Merchants reported a 28 percent chargeback rate on AI-driven orders, faced unauthorised dropshipping logistics, received anonymised customer emails that made returns impossible, and had confidential wholesale pricing publicly exposed. Some brands were accused of violating 'no-Amazon' clauses in their wholesale contracts and suffered reputational damage when Amazon's AI misaligned their branding or return policies.
Reported impact
For directly impacted merchants, the incident signals a loss of control over their own brand identity and inventory. They are being forced into a 'platform-as-intermediary' model where they bear the operational risk of shipping, returns, and complaints while Amazon controls the customer relationship.
Critics argue that as AI agents become the primary interface for shopping, brand loyalty may be replaced by algorithmic decisions, and current digital laws are insufficient to protect data from purchasing agents that act as proxies for users.
Index analysis
The incident highlights a double standard in corporate transparency. Amazon recently sued AI startup Perplexity for scraping Amazon's own site, yet it deployed similar scraping technology against small businesses as an 'experiment.'
Amazon designed the trial as an automatic enrollment programme, placing the burden on business owners (who did not know the programme existed) to email a specific address to be removed. The AI agents operate as 'black boxes,' with merchants not notified when their data was scraped and lacking tools to verify how the AI modified their product information before presentation to millions of shoppers.
Governance controls
Amazon's approach prioritised speed and scale over established norms of merchant autonomy and data ownership, reportedly driven by a desire to capture 'customer intent' outside its own ecosystem.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.