{
  "id": "B15",
  "url": "https://cartforagents.com/guides/ai-shopping-personalization",
  "title": "AI shopping personalization: control context before comparing results",
  "description": "Compare thin and richly constrained buyer profiles, record memory and country settings, and preserve product disqualifiers in a context-control card.",
  "answer": "OpenAI, Google and Perplexity all say shopping results can depend on the shopper: memory, custom instructions, search history, location, language and what the shopper says in the conversation. So a single screenshot is one shopper's result, not your rank. Before comparing results, record the account, memory and personalization settings, country and language, and run matched thin and richly constrained versions of the same request, keeping any facts that should rule your product out.",
  "take": "Personalization is good news for specific products. The more a shopper tells an assistant, the better a listing with exact facts can match them: in our simulation, detailed shopper briefs found the factual listing every time on GPT and both Gemini models.",
  "summary": "Two people can type the same words into the same assistant and see different products, and the platforms say so in their own help pages. Before you compare any result, write down whose result it was.",
  "claims": [
    {
      "text": "OpenAI says a product appears in ChatGPT's shopping carousel when ChatGPT perceives it as relevant to the shopper's intent, considering the query and context such as Memory or Custom instructions; that a stated budget makes it focus more on price; and that it may leave out products the shopper has said they dislike.",
      "source_title": "OpenAI Help Center: Shopping with ChatGPT Search, How product results are selected",
      "source_url": "https://help.openai.com/en/articles/11128490-shopping-with-chatgpt-search",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Living help article; lists inputs, not weights. It also says the model can make mistakes interpreting intent."
    },
    {
      "text": "OpenAI says ChatGPT memory can use past chats, saved memories and custom instructions, and, depending on plan and region, relevant Library files and connected-app content when those features are available and enabled; it says memory features and controls vary by plan, region, platform and workspace, with controls under Settings > Personalization > Memory.",
      "source_title": "OpenAI Help Center: Memory in ChatGPT",
      "source_url": "https://help.openai.com/en/articles/8590148-memory-faq",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Describes personalization in general, not shopping specifically."
    },
    {
      "text": "OpenAI says ChatGPT may generate simplified product titles and descriptions, based on information it receives from third-party providers, to make results easier to read.",
      "source_title": "OpenAI Help Center: Shopping with ChatGPT Search, Product descriptions, labels, reviews, and ratings",
      "source_url": "https://help.openai.com/en/articles/11128490-shopping-with-chatgpt-search",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Does not say which facts are kept when a description is shortened."
    },
    {
      "text": "OpenAI's merchant page says shopping is currently live for ChatGPT users in the U.S. and that ChatGPT helps match products to people based on shopper preferences and past interactions.",
      "source_title": "OpenAI: Power product discovery in ChatGPT",
      "source_url": "https://chatgpt.com/merchants/",
      "source_date": "2026-10-27",
      "method": "vendor-claim",
      "limitation": "A merchant marketing page. The help center separately says shopping is available on mobile and web without naming a country."
    },
    {
      "text": "Google says Personal Intelligence in AI Mode references previous searches and activity saved in Search Services History, is available to users 18 or older with history and personalized recommendations enabled, and that memory in AI Mode is available in the US, in English.",
      "source_title": "Google Search Help: Personal Intelligence, How AI Mode in Search personalizes responses for you",
      "source_url": "https://support.google.com/websearch/answer/17212611?hl=en",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "The page warns that Search services settings are being updated and may not match the current experience."
    },
    {
      "text": "Google says that with Personalized Recommendations off, Search services still use current context such as location, language, device type and recent searches, and that results can vary between people for reasons other than personalization, such as language settings or localized results.",
      "source_title": "Google Search Help: Manage Personalized Recommendations in Search services",
      "source_url": "https://support.google.com/websearch/answer/17026260?hl=en",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Settings are rolling out gradually; the page says older accounts may still use Web & App Activity."
    },
    {
      "text": "Google Shopping Help says products are ranked on relevance, search terms and other Google activity, and that the products a shopper sees may be based on their searches, views and other browsing on Google and on saved shopping preferences.",
      "source_title": "Google Shopping Help: Understand how shopping results are generated, Personalization",
      "source_url": "https://support.google.com/googleshopping/answer/15456476?hl=en",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Google Shopping; we do not assume every AI Mode or Gemini answer follows the same rules."
    },
    {
      "text": "Perplexity's November 25, 2025 launch post says its shopping experience remembers past searches and learns a shopper's patterns, keeps context across follow-up questions, and launched free for US users.",
      "source_title": "Perplexity: Shopping that puts you first",
      "source_url": "https://www.perplexity.ai/hub/blog/shopping-that-puts-you-first",
      "source_date": "2025-11-25",
      "method": "vendor-claim",
      "limitation": "A launch announcement, not current documentation. The merchant program terms URL listed in our research returned \"Page Not Found\" on 2026-10-27."
    },
    {
      "text": "Shopify's Global Catalog documentation says agents can pass buyer signals for relevance and localization (country, region, postal code, language, currency and intent) and a buyer-linked token for personalized search results for a signed-in Shop customer.",
      "source_title": "Shopify: Global Catalog MCP, Personalized search and search parameters",
      "source_url": "https://shopify.dev/docs/agents/catalog/global-catalog",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Describes what an agent can send to the catalog, not what any assistant actually sends."
    },
    {
      "text": "OpenAI's product feed specification says the standard upload currently targets the US, that row-level market columns do not change that default, and that an omitted target country does not mean worldwide.",
      "source_title": "OpenAI: Product feed specification, Geo tagging",
      "source_url": "https://developers.openai.com/commerce/specs/file-upload/products",
      "source_date": "2026-10-27",
      "method": "documented",
      "limitation": "Merchant-side feed rules; says nothing about shopper personalization."
    },
    {
      "text": "In a supplied-snippet simulation, an invented humidifier with its facts as a spec line was picked by GPT-6.1-Sol, Gemini 3.8 Flash and Gemini 3.1 Pro in every session with a detailed shopper brief (20 of 20, 19 of 19, 15 of 15) and in none with a one-line ask or no shopper (0 of 40 on each model); three Claude models picked it in 59, 32 and 43 of 60 sessions for the detailed brief, one-line ask and no shopper.",
      "source_title": "Cart for Agents: Vague vs specific product facts, simulated picks by Claude, GPT and Gemini models",
      "source_url": "https://cartforagents.com/research/vague-specific-product-facts-simulation",
      "source_date": "2026-10-27",
      "method": "controlled-simulation",
      "limitation": "Models chose from snippets we supplied; the product was invented and the detailed brief was written to fit it. API models, not ChatGPT, Gemini or AI Mode shopping. Each model picked 2 of 8, so random picking is 25%."
    },
    {
      "text": "In the same simulation, with a one-line ask, Gemini 3.8 Flash picked the vague humidifier listing in 5 of 20 sessions and the factual one in 0 of 20.",
      "source_title": "Cart for Agents: Vague vs specific product facts, simulated picks by Claude, GPT and Gemini models",
      "source_url": "https://cartforagents.com/research/vague-specific-product-facts-simulation",
      "source_date": "2026-10-27",
      "method": "controlled-simulation",
      "limitation": "Small cells with overlapping intervals; the vague copy repeated the shopper's own words."
    },
    {
      "text": "Our 162-search Shopify Catalog snapshot of October 26, 2026 used the default request with no country, currency, postal code or buyer token; it returned 1,620 result occurrences, one variant per product, and the native-checkout flag was false on all 1,620 returned variants.",
      "source_title": "Cart for Agents: Native-checkout flags, interpret a Catalog snapshot without overclaiming",
      "source_url": "https://cartforagents.com/research/ucp-catalog-native-checkout-research",
      "source_date": "2026-10-26",
      "method": "observed",
      "limitation": "One unpersonalized context. Occurrences, not distinct products. Not a ChatGPT result."
    }
  ],
  "sources": [
    {
      "title": "OpenAI Help Center: Shopping with ChatGPT Search",
      "url": "https://help.openai.com/en/articles/11128490-shopping-with-chatgpt-search",
      "accessed": "2026-10-27"
    },
    {
      "title": "OpenAI Help Center: Memory in ChatGPT",
      "url": "https://help.openai.com/en/articles/8590148-memory-faq",
      "accessed": "2026-10-27"
    },
    {
      "title": "OpenAI: Power product discovery in ChatGPT",
      "url": "https://chatgpt.com/merchants/",
      "accessed": "2026-10-27"
    },
    {
      "title": "OpenAI: Product feed specification",
      "url": "https://developers.openai.com/commerce/specs/file-upload/products",
      "accessed": "2026-10-27"
    },
    {
      "title": "Google Search Help: Personal Intelligence, How AI Mode in Search personalizes responses for you",
      "url": "https://support.google.com/websearch/answer/17212611?hl=en",
      "accessed": "2026-10-27"
    },
    {
      "title": "Google Search Help: Manage Personalized Recommendations in Search services",
      "url": "https://support.google.com/websearch/answer/17026260?hl=en",
      "accessed": "2026-10-27"
    },
    {
      "title": "Google Shopping Help: Understand how shopping results are generated",
      "url": "https://support.google.com/googleshopping/answer/15456476?hl=en",
      "accessed": "2026-10-27"
    },
    {
      "title": "Google Shopping Help: Recommendations & personalization",
      "url": "https://support.google.com/googleshopping/answer/9116422?hl=en",
      "accessed": "2026-10-27"
    },
    {
      "title": "Perplexity: Shopping that puts you first",
      "url": "https://www.perplexity.ai/hub/blog/shopping-that-puts-you-first",
      "accessed": "2026-10-27"
    },
    {
      "title": "Perplexity: Shop like a Pro",
      "url": "https://www.perplexity.ai/hub/blog/shop-like-a-pro",
      "accessed": "2026-10-27"
    },
    {
      "title": "Shopify: Global Catalog MCP",
      "url": "https://shopify.dev/docs/agents/catalog/global-catalog",
      "accessed": "2026-10-27"
    },
    {
      "title": "Cart for Agents: Vague vs specific product facts, simulated picks by Claude, GPT and Gemini models",
      "url": "https://cartforagents.com/research/vague-specific-product-facts-simulation",
      "accessed": "2026-10-27"
    },
    {
      "title": "Cart for Agents: Native-checkout flags, interpret a Catalog snapshot without overclaiming",
      "url": "https://cartforagents.com/research/ucp-catalog-native-checkout-research",
      "accessed": "2026-10-27"
    },
    {
      "title": "Cart for Agents: Product constraint coverage",
      "url": "https://cartforagents.com/guides/product-constraint-coverage",
      "accessed": "2026-10-27"
    }
  ],
  "limitations": [
    "Checked on 2026-10-27. Personalization settings are changing; Google's pages say its Search services settings are being updated. Re-check before acting.",
    "The platforms document which inputs they consider, not how much each one weighs. We cannot see what any assistant shows any shopper.",
    "Perplexity's statements come from launch posts, not current documentation, and the merchant terms URL we had on file returned \"Page Not Found\".",
    "The context-control card is our proposal. It is not a standard, and we have not measured how much each setting changes results on any live platform.",
    "The thin-versus-rich evidence is a supplied-snippet simulation with an invented product and API models. It shows that context can change picks under controlled conditions, not by how much on ChatGPT, Gemini or AI Mode.",
    "No brief in that simulation asked for something the product could not do, so whether stated limits help an agent turn away the wrong shopper is untested."
  ],
  "related": [
    "https://cartforagents.com/guides/how-ai-shopping-agents-choose-products",
    "https://cartforagents.com/guides/ai-shopping-platforms",
    "https://cartforagents.com/guides/how-to-rank-in-chatgpt",
    "https://cartforagents.com/guides/ai-visibility-audit",
    "https://cartforagents.com/guides/ai-shopping-prompt-tracking",
    "https://cartforagents.com/research/vague-specific-product-facts-simulation",
    "https://cartforagents.com/guides/negative-product-attributes"
  ],
  "published": "2026-10-27",
  "updated": "2026-10-27",
  "checked": "2026-10-27",
  "kind": "guide"
}