The Myth That Product Copy Is Dead
You've heard the take. AI shopping agents read your product feed and your schema, so why bother writing product descriptions anymore? Just get the structured data clean and let the robots do the rest.
It sounds efficient. It's also one of the most expensive misconceptions floating around ecommerce right now.
Here's the blunt truth: feeds, schema, and agentic commerce are one source of traffic and conversions, and they cannot access everything from a product feed or schema. AI is still in its infancy. It will grow. But it's not something that can carry a business on its own today. Write your pages for the humans who land on them, and the machines catch more than you'd think.
AI Is One Channel, Not The Whole Funnel
A lot of brands are so busy chasing the new shiny thing that they're gutting the pages that actually pay the bills. Cover your brand on the product page instead of hollowing it out because AI is trendy. That's a mistake multiple companies are making right now.
Think about it like TikTok. The platform lets people check out without ever leaving the feed, not unlike the ChatGPT and Shopify deal. That didn't kill websites or apps. People still go to search engines to find products and reviews on YouTube. They subscribe to email and SMS lists for deals. They visit sites directly, through affiliate links, or through PPC. If your product page doesn't share the solutions your product actually provides, those people drift somewhere else.
I'll go personal on this. I researched my favorite travel jacket on the SCOTTeVEST website, watched the YouTube reviews, liked what I saw, and still didn't trust the site itself enough to buy direct. So I bought it on Amazon instead. That tells you something. Their page was good enough to convince me on the merits but not good enough on trust or return policy to close me. Now imagine that page was also out of date and missing details. I'd have walked to a competitor entirely, and SCOTTeVEST would have lost the sale completely.
People Buy When They Feel Understood
Customers don't pay for a spec sheet. They pay because they feel like you get them, and good product copy proves it in seconds.
In my own shopping research I started with five companies. Three of them had lackluster website experiences, which made it painfully easy to narrow it down to two. Those three felt like knock-off operations. They skipped the detail that mattered, like whether there's a padded space for an iPad, and there was nothing human about the words on the page. Gone.
A great product page answers the buyer's situation: will these running shoes work specifically for pronation and trails? Is this costume too large for a Chihuahua but right for a French Bulldog? Get specific. That's the difference between a tab they close and an order they place. The same specificity that convinces a human is what helps a model cite you, which is exactly the point of Specificity Is the New Keyword: How Narrow, Focused Content Wins AI Citations.
Google doesn't care how many hours you sweated over a product page. You either sell the solution in a few seconds or the visitor bounces. So sell.
And here's the part people miss: AI crawling your website sees the same content the person sees. A page that's clear to a shopper is also a page where the AI learns more about your product, which means it can better match that product to the right query. You don't have to write two versions. Good copy does double duty.
The Feed Can't Hold Your Whole Story
Feeds have character limits. Your webpage does not.
That's not a small distinction. Your page is where you tell the brand story. Where you describe how the product is made, why the materials matter, how it fits into the buyer's actual life. Where you differentiate yourself from the lookalike sitting right next to you in the comparison list. Detail gets lost in a feed because the feed doesn't allow for that much text. Your page does.
It works the other direction too, and this is where it gets interesting. Your influencers, ambassadors, affiliates, ecommerce editors, and hired creators don't get to call you every time they have a question. When they're ready to produce content, they research your products by reading your product pages. And if those creators end up being a source an AI pulls from, that's exactly how the model discovers the same talking points it found on your page and hopefully in your feed.
When all the trusted sources start saying the same thing, the model makes a reasonable assumption: this product and this page are the right ones to recommend. Whether the conversion happens on your site or inside the LLM, you got the sale instead of a competitor. If the selling points and the user experience aren't on your product pages, the people making content about your brand may never mention them, and you've quietly deleted a learning point the models needed.
But Don't Neglect The Machine Side
Now the honest caveat. Clean human copy alone won't get you into agentic recommendations either, and the bar for the data is higher than most retailers expect.
Agentic commerce needs an accurate, up-to-date catalog with detailed schema on every product, plus fast, reliable responses and near real-time inventory. Google's documentation on its agentic protocol calls out three signals that decide whether an AI even considers your product: price validity, delivery time, and the merchant return window. If those fields are missing or incomplete, your product isn't ranked lower. It just isn't included. Period.
A recent audit of 500 top retail sites found 65.8% weren't sending structured data via JSON-LD, 26.6% were missing return policy schema, and 17.6% were missing price validity. That's a shocking share of the biggest names in retail effectively invisible to agents. A product detail page can be a beautiful piece of on-page SEO, and if the agent can't read the structured data, none of the pretty helps. If you want a practical view of how that markup shapes what AI engines say about your products, start with How Structured Data Trains AI Snippets and Grows Your Search Visibility.
Do Both, Properly
Stop treating this like an either-or. Your product page and your feed both need to carry the details that matter, just in different formats. The feed needs structured data for machines. The page needs rich, human descriptions for the person holding their wallet and for the AI crawling what they see.
Practically, audit your top sellers for the selection signals your target platforms actually read, set a realistic number of products your team can optimize each month, and get your feeds updating as close to real time as you can manage so a transaction never fails on stale inventory. If you're going to feed AI traffic at all, it needs to be as accurate and current as possible. That usually means organizational change too, which is why Who Should Own the Product Feed? Why SEO Must Partner With Paid in the AI Era is worth a read before you assign this work.
If you want to go deeper on the machine-readable side, two of our own pieces cover it well: AI Shopping Isn't Reading Your Website—It's Reading Your Data and AI Can't Recommend What It Can't Understand: How to Make Your Products Easier for AI Systems to Evaluate.
The winners in this next era won't be the brands with the cleanest feed or the prettiest page. They'll be the ones who got both right and wrote copy a human actually wanted to buy from.
Sources
This piece draws on the Ask An SEO column asking whether product page copy matters since agents read feeds and schema and on the Search Engine Journal report that 70% of top retailers are invisible to agentic commerce.
