Walmart Opens Checkout to AI Shopping Agents as Amazon Blocks Them

What happened

Walmart has introduced technology that lets AI agents make purchases on customers' behalf, with controls over spending and approval of transactions, according to a TechCrunch report published on 21 October 2026. The move contrasts with Amazon's decision to block AI agents from shopping on its site.

The divergence between the two retailers is the central development. One large retailer is opening its checkout to autonomous shopping software, while another is restricting it. The reporting presents this as a choice between two approaches to agent-led commerce, not an agreed industry standard.

The available reporting is a high-level summary, so the rollout's scope, availability and partners are not confirmed here. The spending-control and transaction-approval features should be read as described in the report, not as a full product specification.

Why it matters

AI agents are moving from answering questions to completing transactions. Once software can buy on a customer's behalf, the purchase journey is no longer only a sequence of screens a person navigates. It becomes a permission structure. Retailers have to decide what an agent may spend, which purchases need a human to confirm them, and how the customer can see and reverse what the agent has done. Those are service-design questions as much as technical ones.

For leaders in experience and digital transformation, the significance lies in the operating model. Spending limits and approval steps are the mechanisms that make delegated buying trustworthy, and they shape whether customers feel in control or displaced. Retailers that block agents and those that admit them will both need to define the trust layer that sits between the customer and the software acting for them.

The Renascence take

Most coverage of this story will focus on the rivalry between two retail giants. The more useful question is how a customer's consent gets built into an agent's workflow.

The real battleground is not whether agents can shop, but whether customers can trust them to do so. Trust is built through visible limits, predictable defaults and easy reversal, and people tend to over-trust automation until the first unexpected charge arrives. A customer-obsessed operator should map every decision point an agent can trigger, set conservative default spending caps, and design approval moments so they feel like the customer's own choice rather than a hurdle. Treat the approval screen as the most important page in the journey, not an interruption to it.