ChatGPT 'Shopping' Is Mostly Google Shopping — What That Means for Your Store

ChatGPT's product carousel is largely populated from Google Shopping's organic index. So 'optimizing for ChatGPT shopping' is really Google Merchant Center + Shopping rank — and one feed won't win all five AI engines.

By Andrej Ruckij · · 6 min read

ChatGPT ‘Shopping’ Is Mostly Google Shopping — What That Means for Your Store

By Andrej Ruckij · June 16, 2026

TL;DR: When ChatGPT shows a product carousel, those picks are largely pulled from Google Shopping’s organic index — not from a private OpenAI–merchant pipeline. Three independent investigators converged on this in 2025–26; one study put it at 83% of carousel products appearing in Google’s top 40 Shopping results. The practical consequence: “ChatGPT shopping optimization” is roughly 80% Google Merchant Center + Google Shopping rank. There is no separate ChatGPT-shopping discipline to learn. And the dependency is ChatGPT-specific — Gemini reads your Merchant Center feed directly, Perplexity and Copilot lean on Bing, Amazon Rufus uses Seller Central. One feed won’t win five engines.

Every week another post announces “ChatGPT shopping” as a brand-new channel you need a brand-new strategy for. The more useful finding is the opposite: when you look under the hood, ChatGPT’s product carousel is mostly Google Shopping wearing a trench coat. That’s good news — it means the work is familiar — but only if you know which feed actually feeds which engine.

What investigators actually found

Three people pulled ChatGPT’s shopping carousel apart, separately, and landed in the same place:

  • Alexis Rylko (Nov 2025) — the earliest and cleanest independent teardown. ChatGPT’s carousel product data (name, image, rating, review count, price, even Google’s srsltid URL markup) was identical to Google Shopping, and he spotted base64-encoded Google Shopping query parameters in ChatGPT’s outbound requests.
  • Semrush (Jan 2026) — replicated it with a controlled test (100 prompts × 5 runs): ChatGPT’s top product appeared in Google Shopping’s first 3 results about 75% of the time, and confirmed the same mechanism via browser dev-tools.
  • Tom Wells / Peec AI, in Search Engine Land (Mar 2026) — the quantification: across 40,000+ carousel products in 10 verticals, 83% were found in Google’s top 40 Shopping results and 60% in the top 10 — versus roughly 11% from Bing.

The mechanism, stated carefully: the path runs Google Merchant Center → Google Shopping’s organic index → ChatGPT. It is not a direct, private “Merchant Center feeds OpenAI” pipe. Your feed matters because it ranks in Google Shopping, and ChatGPT reads Google Shopping.

How solid is this? (the part most posts skip)

Here’s the honesty that the “83%!” headlines leave out:

  • The thesis — ChatGPT shopping piggybacks on Google Shopping organic — is genuinely triangulated. Three independent methods, same answer. Trust it.
  • The exact 83/60 numbers are single-source (Peec AI). The independent replication measured a different-but-consistent figure (Semrush: ~75% in the top 3). So cite “one study found ~80%,” not “research proves 83%.”
  • A conflict-of-interest worth naming: Search Engine Land is owned by Semrush. The SEL article and Semrush’s own study aren’t fully independent of each other. The clean independent corroboration is Rylko’s teardown.

None of that weakens the takeaway. It means you can act on the direction with confidence and treat the decimal points as provisional — which is exactly the right posture in a space this young.

The consequence: there’s no new discipline to learn

If 80% of ChatGPT’s picks come from Google Shopping’s organic index, then optimizing for ChatGPT shopping is optimizing for Google Shopping — clean Merchant Center feed, accurate attributes, competitive price, strong Shopping rank. The skills transfer directly. Nobody needs to buy a “ChatGPT shopping course.”

But one feed won’t win five engines

The trap is assuming what works for ChatGPT works everywhere. It doesn’t — each AI shopping surface pulls product data from a different place:

EngineWhere its product picks come from (mid-2026)
ChatGPTGoogle Shopping’s organic index (crawl fallback) + its own feed program for checkout
GeminiYour Google Merchant Center feed, directly (Google-native)
PerplexityBing index + own crawl + external signals (Reddit, YouTube, reviews)
CopilotBing index
Amazon RufusYour Seller Central listings

So the real multi-engine play is: win Google Shopping (covers ChatGPT and Gemini), keep Bing Merchant Center healthy (Perplexity, Copilot), and treat Amazon as its own world. That’s three feeds, not one — and not five separate disciplines either.

The feed-vs-crawl nuance

ChatGPT shopping is actually both feed- and crawl-driven, and the balance is shifting. OpenAI runs an official feed-based merchant program (Shopify and Etsy auto-sync into it), but that path powered Instant Checkout — which OpenAI wound down in March 2026. The discovery carousel is the part that falls back to crawling Google Shopping, and discovery is where most of the action is. Expect the mechanics to keep moving.

What to do now

  1. Treat your Google Merchant Center feed as the master asset. Accurate titles (lead with the terms buyers use), GTIN/MPN, brand, price, availability, shipping/returns. This one feed reaches both ChatGPT and Gemini.
  2. Earn Google Shopping organic rank, not just paid placement — the carousel reads the organic index.
  3. Keep Bing’s Merchant Center alive so Perplexity and Copilot can see you.
  4. Don’t build a “ChatGPT shopping” project. Build feed quality and Shopping rank; ChatGPT visibility follows.
  5. Re-test after model and platform updates. This is the fastest-moving corner of commerce right now.

Honest caveat

The dated facts here are accurate as of June 2026 and will move — OpenAI changed its shopping approach twice in six months. What’s unlikely to change soon is the structural point: AI shopping engines lean on the established product indexes (Google Shopping, Bing, Seller Central) rather than inventing private ones. Optimize the indexes, not the chatbot.

Key takeaways

  • ChatGPT’s product carousel is largely populated from Google Shopping’s organic index — triangulated across three independent investigations (the exact 83/60 figures are single-source; the thesis is not).
  • “ChatGPT shopping optimization” ≈ Google Merchant Center feed quality + Google Shopping rank. No separate discipline.
  • One feed won’t win five engines: Gemini = Merchant Center, Perplexity/Copilot = Bing, Rufus = Seller Central. Win Google Shopping, keep Bing healthy, treat Amazon separately.
  • ChatGPT shopping is both feed- and crawl-driven; Instant Checkout was wound down in March 2026 and discovery now leans on the Google Shopping crawl.
  • Act on the direction with confidence; treat the specific percentages — and the mechanics — as provisional.

Sources