Sep 16, 2026 · 42 min
AI agents challenge Amazon’s grip on shopping
AI Wants to Shop for You. What Could Go Wrong?
Delegating purchases to software could change who controls consumer relationships, payments, advertising, and the economics of retail.
- 1Agentic shopping promises convenience, but unreliable data, restricted inventory access, and uncertain business models limit autonomous purchasing.
- 2Trust becomes the central barrier when agents need access to inboxes, logins, payment cards, and expensive buying decisions.
- 3Amazon has powerful defenses, yet Walmart and Target may partner with AI companies to weaken its advantage.
Don't miss
The hosts connect agentic commerce to Amazon’s greatest vulnerability: losing the customer relationship while retaining only fulfillment.
The brief
Jason Del Rey joins Peter Kafka to define agentic commerce: software that moves beyond product suggestions toward researching, deciding, and completing purchases for consumers.
The promise is clearest for repetitive shopping, especially groceries, but the systems still face incomplete data, restricted inventory access, and unclear reasons for consumers to delegate.
Trust is the hard test. Early agents such as Instinct and Meta’s Muse can reach inboxes and accounts, but failed refunds and account bans show the cost of mistakes.
Amazon could lose the customer relationship if outside agents reduce it to fulfillment, yet its logistics, Prime membership, delivery network, and own AI tools provide substantial defenses.
The deeper question is whether an agent represents the shopper or whoever funds it through advertising, commissions, or transaction fees—a conflict that could reshape retail competition.