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

Zohar Gilad is cofounder and CEO of Fast Simon.

gettyIn a recent survey, 24% of shoppers said they would be comfortable letting an AI agent shop for them. While AI-driven traffic to online stores is still a small share of visits, it’s growing quickly.

These numbers will undoubtedly increase. So what happens when AI becomes the intermediary between a brand’s catalog and the shopper?

Because the technology is developing so rapidly, anyone claiming to have a definitive answer or playbook on merchandising for AI is suspect. The more useful exercise is to ask harder questions about what happens to product discovery when an AI system starts deciding which products a shopper sees.

Here are four issues that AI-savvy retailers should be thinking about now.

Traditional SEO taught marketers to compete for visibility inside a ranking system: Identify search intent, optimize pages, earn authority and improve your position.

Now, a new system has emerged: AI (or generative) engine optimization (AEO/GEO).

Unlike keywords, AI prompts are essentially infinite and personalized, and countless different prompts can be clustered around the exact same consumer intent. Consider these requests:

• Find me a breathable, floral dress for under $200.

• Give me a romantic, cottage-core dress appropriate for a guest at a summer wedding.

• I’m going to an outdoor wedding next month, and it’s going to be 85 degrees; what dress will look elevated but won’t show sweat?

Each prompt could lead AI to recommend the same product or three different ones. With open-ended prompts, there’s little predictability.

The answer isn’t to optimize for every possibility, but to make the underlying product information understandable to AI. Then it can connect a product to the many different ways a shopper might describe it (fabric, weight, seasonality, occasion, etc.).

That creates a new merchandising question: How do you structure your catalog to map to these infinite systems? And when the system deciding what gets surfaced is contextual and constantly changing, how do you reliably earn product visibility?

AI agents are well-suited for the shopping that consumers view as bothersome chores: groceries, household essentials and even the frustrating process of finding a holiday gift for Aunt Mathilda.

Conversely, shopping for high-consideration items, like a designer handbag, involves a desire to dream, be inspired and browse. Many shoppers actively enjoy this process and wouldn’t want to hand it off to an AI agent.

That prompts the question: Which parts of the shopping experience would a consumer actually want to delegate?

AI’s impact will likely be most immediate in the “chores” space. Merchandisers selling commodities should consider what happens when AI can compare products on price or availability without exposing the consumer to the brand experience.

Which parts of your catalog become commodities once an AI agent strips away the marketing and compares products side by side? Even for commoditized items, how will descriptions, origin countries and imagery influence the agent’s final selection?

The more important question might be what happens on the consumer’s end. If discovery is part of your value proposition, what does an AI agent remove from the brand experience?

Traditional search lets consumers scroll through and click on numerous options. But as AI search models try to catch up to human discovery habits, they distill endless results down to three to five recommendations.

When presented with AI’s narrowed choices, humans default to convenience, accepting what AI shares. If ChatGPT presents three options for the best running shoes, the shopper is unlikely to look elsewhere for a fourth option.

So how does a brand secure its place in the coveted golden set of results when AI is significantly limiting consumer choice?

In traditional search, a retailer competes for page-one visibility and relies on the shopper to browse. In AI search, the retailer may need to earn consideration before the shopper ever sees the product.

That could create a new competitive real estate, a small shelf of products that AI deemed worth showing. And if AI systems are increasingly determining which products matter, how will this narrowing of options alter market competition?

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

That creates an uncomfortable question for merchandisers: If AI controls the interface, what happens to the careful curation they’ve built into the catalog?

We’re already moving in this direction. For example, Shopify’s model context protocol (MCP) server displays a brand’s entire catalog to AI agents, leaving all the product discovery power in the hands of AI platforms.

In this instance, a retailer’s products may be more accessible to AI, but accessibility isn’t the same as merchandising.

Merchandising is about deciding which products to emphasize, how to group them and what story to tell around them. An AI agent can potentially bypass much of that curation and make its own decisions about what to show the shopper.

To exert more influence over the shopping process, brands may need to deploy their own AI to interface with the shopper’s AI.

But what will that look like? Will merchants use their own AI to negotiate with or even manipulate the consumer’s shopping agent? How can retailers use technology to restore their product discovery power?

The transition toward merchandising for bots is imminent, but a one-size-fits-all approach will fail.

The first question for retailers is what they’re willing to give up when AI becomes the intermediary between their catalog and the customer.

As AI shopper agents threaten to remove some of the merchandiser’s control, retailers should resist the urge to immediately rewrite their strategies based on unproven theories. The brands poised for success are those that are actively interrogating their catalog’s purpose, data structure and defensive posture against AI intermediaries.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?