How retailers and brands stay relevant when AI agents do the shopping

A wide beach with palm trees and deck chairs, a cool drink in your hand, the sun sinking into the sea. Retailers and brand providers use scenes like these to appeal to people and sell their products. But if agentic AI does the shopping for consumers in the future, how can vendors influence their product selection? And how will customer relationships work with an autonomous AI assistant acting as an intermediary? Here are some initial tips to help you prepare for the disruption caused by agentic commerce.

Researchers from MyCustomAI, Columbia University, and Yale University set out to explore how AI agents make shopping choices in online marketplaces. Their central question was: “What is our AI agent buying – and why?”

2. Is there any detectable bias?

Yes, AI agents seemed to show strong position bias, favoring products higher up but in varying columns – so there was not a universal ‘top rank’ to be identified.

3. Are there differences between different AI models?

Yes. Different AI models make very different product choices, show different biases, and vary in how rational or consistent they are. For example, one model might heavily prioritize ratings, while another favors product placement.

Conclusion: The authors emphasize that “these findings highlight that the rise of AI agents might significantly affect market demand, and also lead platforms to adjust the type of levers they use.“ The results therefore show that there are no universally applicable factors or rules according to which AI agents evaluate the purchase of products. “Motivated by scenarios where sellers themselves use AI agents to optimize product listings, we show that a seller-side agent that makes minor tweaks to product descriptions – targeting AI buyer preferences – can deliver substantial market-share gains if AI-mediated shopping dominates.“ In other words, purchasing decisions made by agentic AI can be influenced, and perhaps even manipulated, by sellers.

As agentic AI starts handling more of the shopping process autonomously, retailers, brands and vendors will run into several challenges.

• Increasing price competition: Agent-based AI can find the lowest prices across all vendors in real time. If price is one of the main criteria for automated purchasing, this creates enormous price pressure.
• Disruption of the product discovery journey: If AI agents take over product research, many other formats of product discovery could unfortunately suffer and lose visibility.
• Loss of direct brand connection: If AI agents mediate most interactions, retailers and brands risk losing customer loyalty and emotional connections with consumers.
• Dependence on AI intermediaries: Retailers may find themselves competing for placement within agent ecosystems (e.g., OpenAI or Google) with even fewer ways to influence visibility than current SEO or ad models allow.
• Loss of access to customer data: Agents may shield consumer data, making it harder for retailers to track behavior, personalize offers, or remarket effectively.

As AI agents take on a larger role in shopping, retailers and brands face a critical question: will these platforms act as partners, gatekeepers, competitors – or all three at once? On the positive side, agents can drive more traffic, streamline the customer journey, and even deliver higher-quality visits with stronger conversion rates and less casual browsing. But there’s also a real risk: AI systems could take control of product discovery, consumer data, and even transactions – shifting customer loyalty from retailers to the agents themselves. These platforms may also operate as black boxes, offering little transparency into customer preferences or decision-making processes. To stay competitive, retailers will need to rethink their approach to data, branding, advertising, customer loyalty, pricing, and content, and adapt their strategies to fit this emerging AI-driven shopping landscape.

As AI shopping agents take over product discovery and purchasing, retailers need to adapt fast. This checklist outlines the key actions – from data infrastructure to brand engagement – to stay competitive in an AI-first marketplace.

Personalized offers & loyalty integration: Provide APIs that let AI agents access individualized bundles, promotions, and loyalty benefits, so they can factor these into purchase decisions on behalf of customers.

Machine-verifiable trust signals: Establish secure frameworks and standardized certifications that allow agents to authenticate data, verify product quality and sourcing, and factor in reputation indicators like reviews, warranties, and performance metrics.

Experiential branding and strong customer relationships: Double down on storytelling, brand communities, and physical experiences where humans engage emotionally – for example, through platforms like Instagram, Pinterest or YouTube, newsletters, live events, stores and pop-up.

In the past decades, retailers and brands have invested a lot in creating emotional connections with consumers, creating a brand image and advertising campaigns with storytelling. Will these measures now come to nothing? With agent-based commerce, shopping becomes a purely transactional activity. After all, agentic AI evaluates cold, hard facts: prices, product features, delivery speed. Emotional images do not appeal to AI, and it also has no loyalty to a familiar brand.

Does that shift the focus solely to the price war? In that case, large platforms with cheap products are likely to come out on top. We are already seeing this trend in some product categories.

How can retailers, specialized suppliers, and platforms counter this? They need to appeal to consumers with their quality, values, and brand name before consumers engage with agentic AI. Even when it comes to simple consumer products, many people rely on brands they trust, designs they like, and colors they recognize. Ideally, retailer and brand names are included in the prompt: “I always buy this from brand XY” or “Check out retailer XY for good deals.”

There are many consumers who value sustainable production, high-quality design, transparency, or regionality. These consumers must either associate these values with a brand, a supplier or a product family. Or, if these criteria are given to the agentic AI for the purchase, it must find them in a machine-readable and convincing manner on the websites and in the products of the respective supplier.

In other words, while agentic AI might seem to strip away emotions and values from the buying journey, these human factors still matter in retail. The challenge for retailers is to translate them into structured, machine-readable signals that AI agents can act on — and to ensure the corresponding offers achieve strong visibility and lasting impact across the broader retail landscape.