Merchandising For Bots: The Critical Questions Retail Leaders Aren't Yet Asking

The rise of AI agents is forcing a fundamental rethink of how products are discovered and sold online.

Zohar Gilad, cofounder and CEO of Fast Simon, notes that while AI-driven traffic is still a small share of visits, it is growing rapidly.

This shift creates a new challenge for retailers: how to merchandise their catalogs for a system that does not use traditional keywords.

Infinite Prompts Vs. Traditional SEO

Traditional search engine optimization (SEO) focused on ranking systems and keywords, but a new system called AI engine optimization (AEO) is emerging.

Unlike keywords, AI prompts are infinite and personalized, with countless variations clustering around the same consumer intent.

For example, a shopper might ask for a breathable, floral dress under $200, a romantic cottage-core dress, or one that won't show sweat at an 85-degree outdoor wedding.

These open-ended prompts make it difficult for retailers to predict which products an AI will surface.

The solution is not to optimize for every possibility, but to make product information understandable to AI so it can connect a product to many different descriptions.

This creates a new question for merchandisers: how to structure a catalog to map to these infinite systems and earn visibility in a constantly changing context.

Shopping As A Chore Or Leisure Activity

AI agents are well-suited for the parts of shopping consumers view as chores, such as groceries and household essentials.

Conversely, shopping for high-consideration items like designer handbags is often a leisure activity where consumers want to dream and be inspired.

This distinction raises the question of which parts of the shopping experience consumers would actually want to delegate to an AI.

For commoditized items, an AI agent can compare products on price and availability without exposing the consumer to the brand experience.

This forces retailers to consider what happens to their catalog when AI strips away marketing and makes items appear as commodities.

Even for these items, descriptions, origin countries, and imagery will influence the agent's final selection.

The Reduction Of Choice

Traditional search lets consumers scroll through numerous options, but AI search models often distill results down to just three to five recommendations.

When presented with a limited set of choices, humans tend to default to convenience and accept what the AI shares.

If ChatGPT presents three options for the best running shoes, a shopper is unlikely to look for a fourth option elsewhere.

This creates a new competitive real estate, a small shelf of products that AI deemed worth showing.

Retailers must now earn consideration before a shopper ever sees a product, as AI systems increasingly determine which products matter.

This narrowing of options could significantly alter market competition.

How To Control Product Discovery

Agentic commerce is making it easier for AI systems to access structured product information and compare products, potentially completing transactions on a shopper's behalf.

This raises an uncomfortable question for merchandisers: if AI controls the interface, what happens to their careful curation?

For example, Shopify's model context protocol (MCP) server displays a brand's entire catalog to AI agents, leaving all product discovery power in the hands of AI platforms.

Accessibility to AI is not the same as merchandising, which involves deciding which products to emphasize and what story to tell.

An AI agent can bypass much of this curation and make its own decisions about what to show the shopper.

To exert more influence, brands may need to deploy their own AI to interface with the shopper's AI.

This could involve using technology to negotiate with or even manipulate the consumer's shopping agent.

Retailers must actively interrogate their catalog's purpose, data structure, and defensive posture against AI intermediaries to succeed.