The Ceramicist Below the Fold

Delia Fontaine-Park does not exist. Neither does Sunday Object, the mid-market DTC kitchen brand where she supposedly runs product. But the problem she describes is real, documented, and currently unfolding across thousands of brands that haven't noticed it yet. We built her from interviews, trade reporting, and emerging research on agent-driven commerce to explore a question real operators are only beginning to ask: what happens when your product is fine, your reviews are strong, your customers are loyal, and you've simply stopped existing in the places where new customers now look?

The dynamics are grounded. AI-driven traffic to US retail sites surged 805% year-over-year by Black Friday 2025.¹ Shopping queries on generative AI platforms grew 4,700% between 2024 and 2025.² And 80% of AI agents don't properly identify themselves when visiting websites³, meaning brands may not even know whether they're being crawled or ignored.

We spoke with Delia over video. She had a ceramic mug from her own product line and held it up at one point to make a point about texture.

You spent a decade in UX before moving to consumer products. How does a design person end up panicking about structured data?

Delia: I didn't come to this through data. I came to it because our acquisition numbers went sideways in Q3 of last year and nobody could explain why. Paid was performing fine. Organic search looked stable. Retention was actually up. But new customer acquisition just softened. Not a cliff. A slope. The kind of thing where every month you go, "huh, that's a little low," and then five months later you realize you've been saying that every month.

We brought in an analytics consultant and she asked a question nobody had thought to ask: "Do you know how you're showing up in agent-driven shopping?" We all just sort of looked at each other.

What did you find?

Delia: We weren't showing up. At all. Not poorly ranked. Absent. I asked ChatGPT to recommend a premium ceramic chef's knife and it gave back five options, none of which were ours, three of which I'd argue are worse products. So I thought, okay, maybe it's the query. Tried twelve different ways of asking. Nothing.

The maddening part? No notification. No delisting email. No flag in any dashboard. We just weren't there. Like a restaurant that's still open but somehow vanished from every map.

Did you eventually figure out why?

Delia: Partially. It was a stack of things. Our robots.txt had been configured in 2023 to block aggressive crawlers, which was totally reasonable at the time, and it turns out that also blocked the crawlers shopping agents use to index products.⁴ So step one: we'd literally locked the door without knowing anyone was knocking.

But even after we fixed that, we were still barely visible. And this is where it gets, I don't know, philosophical? Our product pages are beautiful. They tell a story. There's a paragraph about the ceramicist in Gifu Prefecture who developed the glaze technique. There's a photo of the knife in a hand, mid-slice, with this gorgeous tomato. A human looks at that and wants the knife.

An agent looks at it and sees almost nothing. We had maybe five structured attributes per product. A competitor had twenty-two. Weight, blade angle, hardness rating, handle material, balance point. The agent doesn't deliberate. It picks the product it can read.⁵

So you added more attributes?

Delia: We did. And this is where I started losing sleep. Some attributes are easy: dimensions, weight, materials. Fine. But the things that make our products ours, the design philosophy, the sourcing story, the way the handle feels in your hand? How do you tag that? "Handle feel: considered"? "Sourcing story: yes"?

Google actually put out guidance telling CPG brands to treat product data as "the new packaging" and to tag things like sustainable sourcing explicitly, because an agent searching for that attribute won't find it unless it's structured.⁶ Which makes sense from their side. But from ours it means: if you can't reduce it to a field in a schema, it doesn't exist.

That sounds like it changes what the product actually is.

Delia: Right now, the recommendation is to tag products based on the problems they solve, because shoppers ask agents for help with a need, not a brand name.⁷ And I get it. But our brand was never built around "solves the problem of cutting vegetables." It was built around the experience of using a tool that someone cared about making. I know that sounds precious. But that's not a problem-solution pair. That's a relationship.

If I redesign our entire product presentation around machine legibility, I'm telling a different story. And in consumer products, the story is the product. You change the story, eventually you change the thing.

Are you doing it anyway?

Delia: Yeah. Of course. Shopping queries on AI platforms grew forty-seven hundred percent in a year.² OpenAI says their results are organic, no paid placements.⁸ Which sounds democratic until you realize it means there's no lever to pull. You can't buy your way back to visibility. You either fit the pattern or you don't.

Does that organic model actually help smaller brands?

Delia: In theory. In practice, it helps smaller brands with clean, structured, machine-readable data and extensive third-party review coverage. Which is not the same thing as "smaller brands."

Reddit mentions apparently matter more than your own product copy now.⁹ We have a 4.8-star average from our customers. But our customers leave reviews on our site, not on Reddit. So that satisfaction is invisible to the system. We built a loyal community and it turns out loyalty doesn't crawl well.

What keeps you up at night about this?

Delia: The 52% thing.

Meaning?

Delia: These shopping agents have about a 52% accuracy rate on complex, multi-constraint queries.¹⁰ Roughly half the time, the recommendation is wrong in some way. But the agent still presents it with total confidence. Still picks something. And the user still buys it.

So yeah. That keeps me up.

If you could change one thing about how this system works, what would it be?

Delia: Publish the criteria. Just tell us. OpenAI says they rank on availability, price, quality, whether you're the primary seller, and whether you've enabled Instant Checkout.⁸ Five bullet points. The gap between those five bullet points and actually appearing in results is a canyon, and nobody will tell you what's in it.

We're reverse-engineering a black box with no feedback loop. At least with Google, you could see your ranking change and correlate it to something you did. Here, you're either in or you're out, and the door doesn't have a sign on it.

Sunday Object has since restructured its product data, added thirty-seven new attributes across its catalog, and begun monitoring agent crawl activity through a third-party tool. Delia reports they're now appearing in some agent recommendations.

She says the knife page still has the paragraph about the ceramicist in Gifu. It's below the fold now, under the spec table.

Footnotes

1. World Economic Forum, citing Adobe Analytics data on AI-driven retail traffic through Black Friday 2025. ↩

2. BoF-McKinsey State of Fashion 2026 report. https://www.businessoffashion.com/articles/technology/the-state-of-fashion-2026-report-agentic-generative-ai-shopping-commerce/ ↩ ↩²

3. DataDome/Security Boulevard research on AI agent identification practices. ↩

4. AthenaHQ and OpenAI guidance on OAI-SearchBot crawler access. https://athenahq.ai/how-to/get-your-product-discovered-chatgpt/ ↩

5. Clickvoyant, "AI Shopping Agents: The Strategic Guide to Agentic Commerce in 2026." https://clickvoyant.com/ai-shopping-agents-the-strategic-guide-to-agentic-commerce-in-2026/ ↩

6. FoodNavigator-USA, citing Google guidance for CPG brands. https://www.foodnavigator-usa.com/Article/2026/02/18/why-cpg-brands-must-prepare-for-ai-shopping-agents/ ↩

7. Axios/SAP, quoting SAP's President & CPO for customer experience. https://www.axios.com/sponsored/ai-is-reshaping-how-we-shop-heres-how-brands-can-keep-up ↩

8. OpenAI, "Introducing shopping research in ChatGPT." https://openai.com/index/chatgpt-shopping-research/ ↩ ↩²

9. Dataslayer, on ChatGPT Shopping's use of third-party review sources including Reddit. https://www.dataslayer.ai/blog/chatgpt-shopping-the-new-discovery-channel-processing-50-million-daily-queries ↩

10. ALM Corp, on ChatGPT Shopping accuracy rates for multi-constraint queries. https://almcorp.com/blog/chatgpt-shopping-research-guide/ ↩