We Taught Robots to Speak Human. It Changes Everything About How Shoppers Find Your Products

Think about the last time you typed something into a search bar. Chances are, you didn't write the way you actually talk. You abbreviated. You keyword-stuffed. You learned, over years of conditioning, to speak the language of machines.

That's over.

The shopper sitting on her couch tonight isn't typing "linen dress size 10 under $300." She's telling an artificial intelligence platform, "I'm going to a wedding in Italy in July, it's going to be hot, and I need something elegant but not too formal. I'm thinking pastel, maybe linen, under $300." And the AI understands her. Completely.

This is the shift that should be keeping every merchant up at night. Not because it's scary, but because the window to get ahead of it is right now.

The Interface Changed. So Did Everything Else

For 30 years, e-commerce was built around a simple architecture: drive traffic to your website, convert visitors, complete the checkout. E-commerce still mirrors the brick-and-mortar shop flow: walk into a store, browse the aisle, purchase. The storefront was the sun that everything else orbited around. However, unlike a brick-and-mortar store the traditional digital storefront is missing one giant thing: context. When you shop for yourself you understand exactly what you’re looking for. That’s the new reality with AI agents.

The current storefront model isn't dead, but it's no longer sufficient.

Discovery is now happening on Gemini, Perplexity, Microsoft Copilot, ChatGPT, and a growing list of AI surfaces where your customers are already spending time. The checkout is increasingly happening there too. Here's the part that changes the merchant's calculus entirely: the front end of the transaction has moved, but the back end still belongs to you. Fulfillment, inventory, customer relationship. You remain the merchant of record. You keep the customer.

But keeping the customer only matters if you can reach them in the first place. In an agentic world, that means showing up accurately in the AI’s answer and making sure the AI can find your products based on a shopper’s specific conversational search. That only happens if your product data is structured correctly.

This Isn't Coming. It's Here

I want to be direct about the timeline because I hear a lot of "we'll get to agentic commerce next year." That window is closing fast.

Pacsun's products are purchasable directly inside Perplexity and Copilot today. Dell has structured a catalog of approximately 7,000 products specifically for AI-driven discovery on OpenAI, covering laptops, desktops, servers, monitors and accessories. As Paul Mansour, global marketing director at Dell, put it: "As AI agents become a more common starting point for product discovery, the quality and structure of product data matter more than ever."

He's right. And the brands that moved fast on data readiness are already seeing it. Merchants that invested in catalog enrichment with Feedonomics saw a 24.8 percent increase in clickthrough rate just from improving the quality of their product data going into the channel. Not from a new ad campaign. Not from a redesigned website. From better data.

Not Every Purchase Will Happen This Way — and That's Fine

Here's something the breathless agentic commerce news tends to miss: product category still matters enormously.

My mascara? I already know what I want. I buy it on a schedule. If I can say "order my mascara, I need it by tomorrow" and an AI handles it, I'm a happy customer. That category is where agentic checkout will move fastest: replenishment, CPG, anything you would describe as a routine buy.

A configurable sofa that needs white-glove delivery? I'm clicking through to the storefront. I want to see it, configure it and maybe talk to someone. The traditional e-commerce storefront and checkout isn't going away for high-consideration purchases. It's just no longer the starting point for discovery.

Merchants need a category-level strategy, not a blanket one. Understand which of your products are thumb-print purchases and which still need the full storefront experience. Start by making sure data is ready for the thumbprint purchase but can support the need for agentic interactions on your branded site for higher consideration items.

The Brain Matters More Than the Channel

Here's the distinction I find myself making constantly. There's the surface: the Gemini app, the Copilot interface, the ChatGPT window. And then there's the brain doing the reasoning behind it. These are not the same thing.

A shopper can ask an AI tool about your products in Slack, in WhatsApp, in a voice interface, in an answer engine you haven't heard of yet. The surface is almost beside the point. What matters is whether the AI powering that answer has been trained on accurate, rich, structured data about your products. Get that right and your products get recommended. Get it wrong and your competitors do.

The job right now isn't to build a presence on every AI channel. It's to make sure your product data is in the brain of whichever AI is powering the answers. It’s the inverse of a traditional marketing strategy; rather than adjust product data for the channel, start with ensuring the product data you own can be discovered, regardless of the channel. That's a data problem, and it's solvable today.

We’re in what I’d call the “get your data ready” phase. The channel partners are building out the shopping experiences. The protocols are being finalized — Google’s Universal Commerce Protocol, which governs how agentic transactions happen securely, is rolling out to merchants now. The infrastructure is ready.

We taught the robots to speak human. The question now is whether your data is ready to answer back.

Al Williams is vice president of market strategy, Commerce, where she helps organizations interpret market shifts and make confident, practical growth decisions.