3 Ways CPGs Can Attract AI Shopping Agents

Attracting artificial intelligence shopping agents boils down to following the “know your customer” maxim while turning it on its head.

You must know your customer. You must also provide the right information in the right ways so your customer can better know you and your products.

That means providing much more detail about your brand, its products and how they’re used than in the past. It also requires that you as a company know yourself — key elements of product planning, production, logistics channels, and beyond — so you can better attract AI shopping agents, deliver on what you promise them, and adapt quickly as changing seasons and tastes bring swings in demand.

Even a spreadsheet-savvy consumer contemplating a high-consideration purchase such a car would be hard pressed to manage, much less digest, dozens or hundreds of ever-changing variables for comparison shopping for the various consumer packaged goods. However, fine-grained, multivariate comparison shopping is a breeze for an AI shopping agent.

While the in-house shopping agents of Ralph Lauren, L’Oréal, and others show promise, the key for most fashion brands and CPGs is to ensure that brand-agnostic AI shopping agents have what they need to recommend your products above those of the competition. Your success will hinge on your data and how your customers (read: AI agents) perceive the data.

AI Shopping Agents Want the Facts

Consider apparel, which is where I specialize. A shopper’s natural-language query could be “Looking for comfy pumps for casual dinners out,” “Find me purple pickleball shoes,” or “Suggest a sustainably sourced black cotton T-shirt for under $30.” We can assume the AI shopping agent knows its master’s brand preferences, price sensitivity, size, fit preferences and, with time, relevant elements of its existing wardrobe based on previous buys as it heads off into the digital wilds on the wings of retrieval augmented generation, or RAG.

The agent won’t swoon for aspirational imagery or florid adjectives. Facts, ones found across the product brand-owned sites, retailers, online marketplaces, and influencer and review sites, are its currency. The facts will differ depending on product and market, but may span price, size, color, materials, ingredients, country of origin, scent or flavor profiles, substitution logic, sustainability information, certifications, accreditations, delivery speed, return policies, user reviews and ratings, and more.

Some of this data is out of a fashion brand or CPG’s hands. However, they do control three important levers in attracting AI shopping agents.

1. Consider your product’s qualities and strengths and standardize.

This is the easiest step. Clearly state product features and strengths with consistent descriptors and quantifiable data across platforms. Given a CPG’s vast and changing product portfolios, harnessing AI tools to generate descriptors and keywords as well as affirm consistency is a must. Such tools exist and will only grow more sophisticated.

Regulatory requirements such as the EU’s Digital Product Passport (DPP) will help foster semantic consistency in some product categories, fashion among them, because the DPP requires structured data on materials and sources, production locations, and environmental factors. AI agents will certainly be accessing DPP data come 2027.

2. Understand product use cases.

Knowing your customer is really about understanding how your customer uses your product. A human will have an intuitive grasp of product usage based on experience and association (i.e., a cleaning product that works on stainless steel sinks will work on stainless steel counters). With AI shopping agents, CPGs must get specific.

Imagery plays a key role here. Australia-based beverage company Lion, maker of XXXX beer and other brands, wanted to understand where and in what contexts people actually consumed its products. The aim was for customers to be able to find their Lion beverage of choice more easily. The company dispatched employees to pubs, events and elsewhere, where they took mobile-phone images of beer and beverage taps. AI embedded in enterprise software seamlessly connected to Lion’s B2B and B2C CX systems, which could then deliver maps to users looking for their favorite drinks nearby.

While Lion wasn’t targeting AI shopping agents, the approach applies here. A human shopper’s query to an AI agent will often include a proposed use case — in the examples above, a casual dinner out and a pickleball game. A shoe company that features situation-specific verbiage or images (which a CPG’s AI tools can translate into verbiage) has a much better shot of attracting an AI shopping agent. In apparel, the old standard of fashion models standing against neutral backgrounds may prove less effective in the world of AI agentic shopping where fast-changing context is everything.

3. Make sure your systems and data foundations are in one place.

You can’t share expected delivery time with an AI shopping agent if your production and logistics systems don’t communicate with your customer-facing or retail/franchise-facing systems that AI shopping agents have access to. And because exactly what AI shopping agents look for remains fluid — and will surely evolve — you need a robust and trusted data foundation that enables adaptability as different data for different products change over time.

Also, unified data lets you adapt quickly if, for example, a boost in sales sparked by success in attracting AI shopping agents brings a spike in returns should AI agent and human opinion diverge. In fashion and apparel in particular, quick feedback is especially crucial given tight seasonal cycles and rapidly changing tastes, especially if there’s a chance to reorder hot products during the season.

There’s no one-size-fits-all approach to attracting AI agents, and fashion brands and CPGs are still experimenting. What’s clear is that standardizing and expanding upon product descriptions, understanding diverse product use cases and explicitly delineating them through know-your-customer language and imagery, and unifying your data will help CPGs sell via AI shopping agents.

Peter Akbar is global vice president and chief customer officer for fashion at SAP.