For a century, marketing operated on a simple assumption: the person who uses a product and the person who buys it are the same individual.

That assumption is breaking.

My new article in the California Management Review examines what I call the Shopper Schism — the structural separation of the consumer from the shopper that occurs when AI agents assume purchasing authority.

The trigger for the research was Amazon’s withdrawal from Google Shopping in July 2025. Analysts called it cost-cutting. I saw something different. Nine days before the withdrawal, Amazon had rolled out its AI shopping assistant Rufus to all U.S. customers. By autumn, Rufus was driving over $10 billion in additional annual sales.

Amazon wasn’t optimising for today’s customer. It was positioning for tomorrow’s. A customer that isn’t human.

When I studied how AI agents actually evaluate products, the findings were uncomfortable:

  • Agents are immune to aspirational imagery and emotional appeals

  • They systematically favour verified specifications over narrative claims

  • A $2M investment in lifestyle photography is functionally invisible to algorithmic assessment

  • Different agents converge on similar choices, potentially reducing market diversity

I interviewed 30 executives. 43% recognised the shift but had no strategy. Less than 3% had formal responses. Over half were still primarily optimising for human shoppers.

The article proposes three strategic frameworks — Data Fortification for major brands, Transparency Arbitrage for challengers, and Experience + API for retailers — along with five immediate actions marketing leaders can take this quarter.

But it also asks harder questions that don’t have clean answers yet: Who do these agents truly serve? What happens to market diversity? What do we lose when browsing and discovery disappear?

This is the research I’ve dedicated my career’s next chapter to, and it’s just the beginning.

Read the full article (open access): https://cmr.berkeley.edu/2026/02/the-shopper-schism-competing-when-ai-agents-become-your-customer/

What’s your organisation doing to prepare for algorithmic customers?