The RealReal updates bot defences to block AI shopping agents

What happened

The RealReal, the secondhand luxury fashion consignment retailer, has said it is updating its bot defence system, a change that will make it harder for AI agents to shop on its website.

The announcement follows a report from The Verge, published days earlier, that an AI agent could be used to take advantage of The RealReal's item reserve system. The retailer's response targets that vulnerability directly, with the aim of keeping automated shopping tools from acting on its inventory.

The available reporting focuses on the decision itself and its trigger. It does not set out the technical details of the update or say how the retailer will distinguish agents from human shoppers.

Why it matters

AI agents that browse and buy on a shopper's behalf are moving from concept to live traffic, and retailers now have to decide how to treat them. This story shows the practical consequence: a feature built for human customers, such as holding an item while someone decides, can be exploited at machine speed. The question for digital teams is less about whether agents exist and more about how automated activity is governed on the storefront.

The stakes are higher for resale and luxury categories, where each item is unique and availability is the product. A reserve mechanism that works well for genuine shoppers but can be gamed by software undermines trust with the people it was designed to serve, and it can make the experience feel unfair to the customers who lose out.

The Renascence take

The instinctive response to an agent that games a system is to lock the door. That reaction is understandable, but it treats the symptom and misses the design flaw underneath.

The real failure here is a reserve system that rewards speed and access rather than genuine intent, so it was always going to be exploited by something faster than a person. Behavioural economics would point to commitment devices: holds that require a real signal of intent, are time-boxed, and carry visible rules, so that whoever or whatever is acting is judged by what it does rather than what it is. Customer-obsessed operators should audit where their scarcity mechanisms reward speed over commitment, and redesign those first before adding more bot walls. Blocking agents may be necessary in the short term, but the lasting fix is a reserve experience that a determined human shopper would find fair and an automated one could not profitably exploit.