How to get products into ChatGPT.

Why citation, not ranking, is the metric

ChatGPT does not produce a ten-blue-link page. It produces a synthesized answer with one to three cited products per shopping query, drawn from a retrieval corpus that overlaps with but is not identical to Google's index. Optimizing for ChatGPT means optimizing for the citation slot, not for a position on a results page. The mechanics differ. The measurement loop differs. The work order differs.

A Shopify brand can hold the number-one Google rank for "best waterproof dry bag" and still be absent from every ChatGPT shopping answer for the same phrase. eCommerce Insights's audits through Q1 2026 found this pattern in roughly one in three catalogs where the team had invested in classical SEO without ever addressing AI-readability. The fix is mechanical, not magical.

Step 1: Allow OpenAI's crawlers

OpenAI publishes three bots that are relevant to a D2C brand. Per OpenAI's bots documentation:

• GPTBot — training crawler. Reads pages for model improvement.
• OAI-SearchBot — search crawler. Indexes pages for ChatGPT search.
• ChatGPT-User — user-initiated browsing. Fires when a ChatGPT user asks the model to fetch a specific URL.

All three should be allowed by default in Shopify's stock robots.txt. Verify by visiting yourstore.com/robots.txt and looking for any explicit Disallow on those user agents. While you are there, publish an llms.txt at the root listing your top product, collection, and policy URLs. The file is a hint, not a requirement, but ChatGPT and Perplexity both honor it as of Q1 2026.

Step 2: Complete Product schema

Every PDP needs valid Product JSON-LD with at minimum: name, brand, sku, image, description, offers (with price, availability, priceCurrency), and aggregateRating where you have real reviews. gtin or mpn are strongly recommended for entity resolution. The schema.org Product type documents the full field list; ChatGPT's retrieval layer is particularly sensitive to brand and aggregateRating for entity clarity.

Most Shopify themes ship partial schema. eCommerce Insights's free product schema generator outputs a complete block for one SKU; the paid product audits the whole catalog and pushes recommended schema to Shopify via metafield. See how to add JSON-LD to a Shopify product for the implementation walkthrough.

Step 3: Rewrite the first 300 characters

ChatGPT's retrieval layer reads top-of-page content with more weight than buried passages. The first 300 characters of a PDP body should answer the buyer's question in plain language and then state three quotable facts. Compare:

Weak opener: "The Wanderer is a beautifully crafted weekend bag designed for the modern traveler who demands quality and style."

Strong opener: "The Wanderer is a 35L waterproof weekend bag for two-night trips. 1000-denier nylon shell, YKK zippers, rated to IPX5, and certified carry-on for Delta and American. Backed by a lifetime repair guarantee."

The strong opener is quotable. The weak opener is not. Quotability drives citation.

Step 4: Add an FAQPage block

Five real questions, five real answers, with visible H3 headings on the page and matching FAQPage JSON-LD in the head. The questions should be phrased the way buyers ask them in ChatGPT — eCommerce Insights's typical prompt set includes "is this waterproof," "does it fit a 17-inch laptop," "what's the return policy," "is this carry-on legal," "how does it compare to [competitor]." Map your FAQ to those questions. Stale FAQs lifted from a CSV ten years ago are worse than no FAQ; they signal neglect.

Step 5: Ground claims in reviews

ChatGPT corroborates brand claims against independent sources. A PDP that says "rated waterproof" with no review citation is weaker than one that says "rated waterproof and verified in 412 customer reviews and a 2025 Outside Magazine field test." Visible review counts, syndicated review-site coverage, and structured Review schema all contribute. The harder it is to find a third-party voice confirming your claim, the lower your citation odds.

Step 6: Run a weekly prompt set

Pick ten purchase-intent prompts a real buyer would type. Run them against ChatGPT once per week. Record per prompt whether your product was CITED (named in the answer), BRAND-only (brand mentioned, product not), or MISSED. The eCommerce Insights SKU Visibility Grader automates this for free; the paid product runs it across your whole catalog and emails a Monday digest. See monitor SKU visibility weekly for the cadence.

Step 7: Iterate on the gap

For each MISSED prompt, add a 150-to-250-word answer block to the relevant PDP that addresses the prompt directly. Re-measure the next week. The lift compounds. Three to four iteration cycles typically move a SKU from 20% to 60% citation rate on its top ten prompts, in eCommerce Insights's experience with mid-market D2C catalogs through Q1 2026. The numbers are directional, not guaranteed.

Ranking on Google is necessary. It is not sufficient. The PDP that ranks is rarely the PDP that gets cited.

Frequently asked questions

Related jobs

Related tools

• ChatGPT product visibility checker — single-product spot check.
• Shopify SKU Visibility Grader — catalog-wide scan, free tier.