SKU Visibility in ChatGPT Shopping: Troubleshooting
Diagnose product and variant visibility in ChatGPT shopping with identity, offer and feed checks. Separate integration errors from a product absent in one result.
How to troubleshoot SKU visibility in ChatGPT shopping
A SKU missing from a ChatGPT-shopping result is an observation, not a diagnosis. First identify the exact product, variant, merchant and surface you expected to see. Then check the supported integration and current source data. Feed acceptance, product eligibility and selection in a particular response are separate stages of the investigation.
Capture the missing-result question precisely
Record the prompt, market, language, time and visible surface. Was the product absent, or was the product present without the intended merchant offer? Use the visibility methodology to preserve a repeatable observation before changing the catalogue.
Do not diagnose a permanent exclusion from one run. Equally, do not dismiss a reproducible data error as normal variation. Keep observation evidence and integration evidence in different columns so the team can investigate each on its merits.
Confirm the supported integration path
Check the integration actually used by the merchant and the account’s current status. OpenAI’s shopping guidance says Shopify Catalog is already integrated; do not assume every Shopify merchant needs a separate custom feed or a PA bridge.
For a direct integration, follow the current OpenAI onboarding documentation. Select the applicable schema before editing data. A file-upload record and an API object are different contracts; copying fields between them can create an invalid payload even when the information looks similar.
Verify identity and variant consistency
The stable file-upload product specification defines its own item and variant-group fields. Map the selected variant consistently across the row, destination, image and offer. Keep identifiers stable through ordinary updates and avoid inventing a GTIN when an authoritative one is unavailable.
For a synthetic mug with blue and green variants, each sellable variant needs a distinct item identity. The shared group identity links them without replacing those item identities. A blue row must not open a green default offer while claiming the blue variant’s price and availability.
Check the offer against the destination
Compare the submitted price, currency, availability and merchant identity with the landing page at the same recorded time. Check that the URL is usable, the relevant image resolves and product descriptions state supported facts. OpenAI’s integration best practices provide the current content and variant guidance.
Classify the issue: invalid record, mismatched variant, stale offer, unsupported integration or no technical error established. Follow the integration’s documented update and status process. Do not promise an immediate display change after a correction, and do not treat an eligibility flag as a ranking instruction.
Close the loop with evidence
Retest the data contract and destination, record the accepted correction, then rerun the same declared observation. A successful repair can establish that the feed and page now agree; it does not guarantee that the product will appear in every answer.
PA’s relevant discussion is the retailer’s catalogue and discovery requirements. Confirm any proposed connector or feed-management responsibility explicitly before presenting it as a supported PA service.
Illustrative example
{"label":"SYNTHETIC, not real offers or production payload","blue_item_id":"demo-mug-blue","green_item_id":"demo-mug-green","shared_group_id":"demo-mug-family","price_each_gbp":30,"blue_availability":"in_stock","green_availability":"out_of_stock","interpretation":"Distinct variants with their own destinations; no appearance or eligibility outcome asserted."}
Sources
- Particular Audience | AI Search, Personalization, Merchandising & Retail Media, checked 2026-10-25
- Shopping with ChatGPT Search, checked 2026-10-25
- Get Started — Agentic Commerce, checked 2026-10-25
- Products — Agentic Commerce file upload, checked 2026-10-25
- Best practices — Agentic Commerce, checked 2026-10-25
How Particular Audience does it
Frequently asked questions
Does an accepted product record guarantee that the SKU appears?
No. Treat acceptance and eligibility as integration states, then record whether the product appears in a defined observation. A response can omit an eligible product. Troubleshoot actual data errors separately from selection in a particular run.
Should every Shopify retailer build a new direct product feed?
No blanket instruction is appropriate. Check the existing integration route and current platform guidance first. Avoid sending conflicting duplicate sources merely because a SKU was absent in one observation. Establish the owner and intended update path before changing the catalogue integration.
What is the first check for a variant mismatch?
Compare the exact item identity, group identity, selected variant, destination, image, price and availability. Determine whether the submitted record and landing page refer to the same sellable option. Fix that inconsistency before interpreting the result as a ranking problem.
Related guides
- How to measure brand visibility in AI shopping: Build a repeatable AI-shopping visibility study with defined prompts, surfaces, identities and denominators. Separate observed presence from traffic and sales.
- What is agentic commerce?: Understand agentic commerce across discovery, comparison and authorised action. Map product data, permissions, checkout boundaries and useful retail measures.