OpenAI’s public documentation supports three limited roles for reviews in ChatGPT shopping: public reviews may inform product results and model-generated summaries, reviews may appear in shopping-research comparisons, and merchant feeds can supply defined review fields. It does not publish a universal review-ranking formula or preferred review threshold.
Last updated: July 24, 2026
Key Takeaways
- Separate public review summaries, shopping-research comparisons, and merchant-feed fields.
- Match reviews, ratings, and counts accurately to the correct product or variant.
- Keep supplied values truthful, internally consistent, and current.
- Do not infer preferred length, count, sentiment, star rating, or recency from undocumented claims.
- Collect and moderate reviews authentically without conditioning incentives on sentiment.
Where can ChatGPT obtain review information?
OpenAI documents public websites and supported merchant data as possible inputs. Its ChatGPT shopping documentation says reviews may be considered for a shopping request and that product results can show model-generated summaries based on reviews from public websites. It also says OpenAI does not verify those reviews and ratings.
The operator action is evidence hygiene: keep public product pages accessible, associate review evidence with the correct product, and distinguish merchant-supplied values from third-party review sources.
How should reviews map to products and variants?
Use stable identifiers and the most specific product relationship available. A review for one size, edition, bundle, or model should not be silently attributed to another when that changes the customer’s understanding. Reconcile review counts and ratings with their source and update them when the underlying data changes.
OpenAI’s product-feed specification defines optional review-count and rating fields and recommends review entries with title, content, and rating values. The specification describes accepted data; it does not state an ideal review length, count, sentiment, or ranking effect.
| Documented context | Available review evidence | Operator control |
|---|---|---|
| Product results | Public review information and summaries | Accessible, accurately associated pages |
| Shopping research | Reviews within comparison information | Current public and merchant facts |
| Merchant product feed | Counts, ratings, and review entries | Truthful fields and product matching |
What does shopping research document about reviews?
OpenAI’s shopping-research documentation says the feature may use merchant product data, public product information, and other retail sources, and that reviews can appear in comparisons. It also notes that changing facts or inaccessible sources can limit results.
That supports a current-source workflow, not an optimization recipe. Assign an owner to product facts, review-source mapping, feed freshness, and contradictions. Preserve the source URL and update date for every aggregate value.
How should reviews be requested and moderated?
Ask customers for honest experience without steering the desired sentiment. Disclose material connections and incentives, apply moderation rules consistently, and do not create, buy, rewrite, or deceptively suppress reviews.
The FTC’s reviews and testimonials rule Q&A addresses fake or false reviews, sentiment-conditioned incentives, undisclosed insider relationships, and deceptive suppression. These are authenticity and compliance controls; they should not be recast as an OpenAI ranking factor.
Which review claims are documented and which are editorial?
Documented: OpenAI may consider public reviews, may show summaries, may include reviews in shopping research, and accepts defined review fields in merchant feeds. Editorial: making review evidence specific, accurately matched, current, and useful helps customers interpret it. Unsupported: any claim that a particular word count, star rating, volume, sentiment, or recency window guarantees visibility.
DGP’s GEO guide, AEO guide, and ChatGPT citation case cover adjacent discovery practices. This page stays within the documented ChatGPT shopping boundary.
Frequently Asked Questions
Does ChatGPT use customer reviews for shopping?
OpenAI says ChatGPT may consider reviews for shopping requests, may summarize public reviews, and may include reviews in shopping-research comparisons.
Does OpenAI verify shopping reviews and ratings?
OpenAI’s shopping documentation says that reviews and ratings shown in product results are not verified by OpenAI.
What review fields can a merchant product feed include?
OpenAI’s current specification defines optional review-count and rating fields and a recommended list of review entries containing title, content, and rating values.
Is there an ideal review count or rating for ChatGPT Shopping?
The cited OpenAI documentation does not state an ideal count, rating, length, sentiment, or recency window, and no such threshold should be inferred from it.
Build an authentic review-evidence system: Get From Zero to Sovereign.
Continue building the review-evidence system with the forthcoming The Social Proof Moat.