Who Owns Shopping Data in the Age of AI?

AI is changing how we shop online and sparking a tug-of-war over who controls the data that powers those experiences.

Big retailers like Amazon, Walmart, and Shopify are taking very different approaches:

• Amazon recently blocked AI platforms from using its data and even pulled its product listings from Google Shopping worldwide.
• Walmart is doing the opposite, allowing AI tools like ChatGPT to access its product information.
• Shopify is somewhere in the middle, as it currently allows AI access but may restrict it if it leads to sales happening off its merchants’ sites.

This battle isn’t just about products. It’s also about customer data like purchase history and shopping preferences. AI platforms, social media companies, and retailers all want to use this information to personalize shopping experiences.

The big question: How can companies share enough data to improve shopping without giving away too much control or revenue? Cloudflare, a major internet infrastructure firm, has recently cracked down on AI scraping of customer websites, and numerous startups are emerging to create paywalls that charge AI platforms for content access. These developments indicate a radical shift in the web’s economic foundations: The traditional 30-year model, where content creators exchanged free access for consumer traffic, is unraveling. Within this upheaval, retailers confront a crucial question about what data to expose and to whom.

The answer is far from clear, as demonstrated by the contrasting approaches of Amazon (restrictive) and Walmart (open).

A likely solution could be a “Commerce Data New Deal” where retailers and AI platforms can balance data access with economic benefit.

Two main types of commerce data are at stake:

1. Product Data: Detailed information about products or services, including descriptions, availability, pricing, and increasingly sophisticated attributes like product usage history, sales rankings, and contextual suitability to specific user groups. While much of this is proprietary to retailers, AI searches can sometimes mine similar data from broader web content.
2. Customer Data: Individual user profiles based on purchase history, demographics, and browsing behaviors. Such data is held not only by retailers like Amazon and Target but also by AI search platforms (ChatGPT, Perplexity) and social media companies (Meta, TikTok).

The ideal shopping experience would integrate all these data layers. Proprietary retailer product and customer data combined with AI-collected web and search data, to offer personalized, highly relevant results. The power of this integration explains the aggressive data scraping by AI platforms and the counter-measures by retailers seeking to preserve leverage amid evolving commerce dynamics.

A promising path forward may involve a middle ground between total data openness and outright blocking. Metered access models could emerge, where retailers selectively share enhanced data feeds, featuring detailed product provenance, ranked sales data, or personalized customer-query results, in exchange for monetary compensation or transaction volume.

For example, an AI could ask Nike for a ranked list of running shoes tailored for competitive female runners or customized recommendations for an individual buyer identified through privacy-preserving hashed identifiers. In both cases, Nike would need to be compensated with either cash, traffic, or a cut of future transactions.

Such controlled data sharing opens a clear opportunity for startups to act as intermediaries, facilitating secure data exchanges, defining monetization frameworks, and dividing transaction rewards equitably.

In short, the old web model, which was free content in exchange for traffic, is breaking down. A new era of commerce data is emerging, where access and control must be balanced with economic incentives.