Noma by Nurdd (NomaHQ)Noma
Sign up

Free playbook · PDF

How to get your products recommended by ChatGPT, Google AI Mode and other AI shopping assistants

A playbook for e-commerce and DTC teams: how ChatGPT, Google AI Mode, Gemini and Perplexity choose products today, a product-attribute worksheet, a structured data and feed checklist verified against vendor documentation, review rules, marketplace and buying-guide strategy, an 84-prompt shopping library and a 90-day plan.

  • Product-attribute worksheet: use case, materials, fit, durability, care, compatibility
  • Structured data and feed checklist for Google Merchant Center and ChatGPT's ACP feed
  • Shopping prompt library: 84 templates across nine categories
  • Review programme plan with the FTC and UK CMA rules built in
  • Shelf share scorecard, citation check and 90-day plan

Get your free copy

Instant download, plus a copy by email. No spam: one email with your link, and a reply from a person if you write back.

The short answer

To get products recommended by ChatGPT and other assistants, give them complete, consistent product facts in three places: page text with Product and Offer structured data, a Merchant Center feed for Google, and an ACP product feed or Shopify Catalog for ChatGPT. Add attributes shoppers filter on, such as use case, materials, fit and compatibility, earn genuine reviews that mention them, and track which products assistants name.

For e-commerce, DTC and retail marketers, merchandisers and SEO leads

What you will learn

  1. Chapter 1

    How assistant shopping works today

    • Shoppers describe a situation to an assistant, and each constraint they state is a filter your product facts must pass.
    • ChatGPT now concentrates on product discovery through ACP feeds and Shopify Catalog, and since March 2026 lets merchants use their own checkout.
    • Google's AI Mode and Gemini draw on the Shopping Graph, fed mainly by Merchant Center; Google says core SEO and feeds matter more than special AI markup.
    • Perplexity and marketplace assistants such as Amazon's read their own catalogues and reviews, so each listing is a product page in its own right.
    • Platform features change month to month: date-stamp your notes and recheck the vendor's page before acting.

  2. Chapter 2

    Product attribute completeness

    • Assistants match shoppers' constraints to product facts, so every attribute a shopper might filter on must be stated as a fact.
    • Seven families cover most shopping questions: use case, materials, fit, durability, care, compatibility and comparisons.
    • Build each category's attribute list from reviews, returns, support tickets and on-site search, not from what your systems happen to hold.
    • Variants need their own identifiers, attributes, images, prices and stock, or the shopper's exact choice cannot be matched.
    • Publish attributes from one source of truth to pages, markup, feeds and listings so every place says the same thing.

  3. Chapter 3

    Product structured data and feeds

    • Markup, feeds and platform integrations are three outputs of one product dataset; keep them in step.
    • Google's merchant listing markup requires name, image and an Offer with price and currency, and recommends shipping, returns, identifiers, attributes and ratings.
    • Merchant Center's conversational attributes, such as questions and answers, manuals and related products, are optional fields aimed at AI Mode.
    • OpenAI's feed requires nine fields per item or variant, and its documentation says recommended attributes such as reviews improve ranking.
    • Checkout on Google's AI surfaces is opt-in per product and needs returns and support details in Merchant Center.

  4. Chapter 4

    Reviews and user-generated content

    • Assistants use review ratings as signals and review text as evidence, so the content of reviews matters as much as the stars.
    • Ask every buyer, at the right time, with questions that draw out use case, fit, durability, care and compatibility.
    • The FTC's 2024 rule bans fake reviews, sentiment-conditioned incentives and review suppression; since April 2025 UK law bans fake and concealed incentivised reviews. Build your programme inside both.
    • Fix the causes of recurring complaints and make the fix visible, rather than trying to bury criticism.
    • Customer Q&As are written as shoppers' questions; answer them well and reuse them in your feeds.

  5. Chapter 5

    Marketplaces, own site, category pages and buying guides

    • Check which domain each assistant cites for your products before deciding where it should be.
    • Marketplaces and your own site are a consistency problem, not a choice: syndicate one attribute dataset everywhere.
    • Category pages earn citations when they explain the choice, compare the range on real attributes and answer common questions.
    • Buying guides built on first-hand testing and honest trade-offs are the kind of content Google says matters in its AI features.
    • Make independent reviewers' jobs easy with accurate facts and unconditional samples, and correct errors politely.

  6. Chapter 6

    Measure your shelf share, and a 90-day plan

    • A frozen, mostly unbranded prompt set by category is the measuring stick for AI shopping visibility.
    • Record treatment, accuracy, destination and cited sources as well as whether your product appears.
    • Shelf share and recommended share per category and engine show where you are missing; accuracy shows which data route is stale.
    • Report observable signals such as assistant referrals, Search Console's generative AI report and survey answers, not invented revenue.
    • Baseline, fix data routes and attributes, then reviews, listings and guides, and rerun at 90 days.

7 worksheets inside

The chapters explain the method. The worksheets are what you fill in to run it: print them, or copy the tables into a document of your own.

  • W1AI shopping surface checklist
  • W2Product-attribute worksheet
  • W3Structured data and feed checklist
  • W4Review and UGC programme plan
  • W5Shopping prompt library: 84 templates in nine categories
  • W6Shelf share scorecard, citation check and 90-day plan
  • W7Run your first baseline
Get the playbook

Questions

How do I get my products recommended by ChatGPT?
Make sure ChatGPT can read complete, accurate facts about each product. Shopify merchants' data reaches ChatGPT through Shopify Catalog; other merchants can apply to send a product feed under OpenAI's Agentic Commerce Protocol. Fill in the recommended fields as well as the required ones, state the attributes shoppers filter on, earn genuine reviews that describe real use, and allow OAI-SearchBot to crawl your pages.
Can I still sell inside ChatGPT with Instant Checkout?
In March 2026 OpenAI said the first version of Instant Checkout did not offer the flexibility it wanted, and that it would let merchants use their own checkout while it focused on product discovery. Merchants wanting deeper integration can build ChatGPT apps. Check OpenAI's current merchant documentation before planning around in-chat checkout.
How does Google AI Mode choose which products to show?
Google says AI Mode draws on its Shopping Graph and on core Search ranking and quality systems. Products reach the Shopping Graph mainly through Merchant Center feeds, supported by Product structured data on your pages. Complete feeds, accurate prices and availability, clear product pages and genuine reviews give it the most to work with.
What is the Universal Commerce Protocol?
The Universal Commerce Protocol (UCP) is an open standard Google launched in January 2026 for agentic commerce, co-developed with retailers and platforms. On Google it powers checkout from eligible retailers inside AI Mode and Gemini, with the retailer remaining seller of record. Products opt in through a Merchant Center attribute, and some categories are not eligible.
Does structured data help products appear in AI answers?
It helps engines read product facts unambiguously, and Google uses it alongside Merchant Center feeds to verify product data. Google also says structured data is not required for its generative AI features and that no special AI markup is needed. Treat Product and Offer markup as part of good product SEO, and keep it identical to what the page shows.
Should I send AI shoppers to my own site or to Amazon?
You rarely control which link an assistant chooses, so first check which domains it cites for your products. Then keep one attribute dataset consistent across your site and every marketplace and retailer listing. Where a marketplace listing is cited ahead of your site, compare the two: the listing may have more reviews, Q&As or clearer facts, which you can match on your own page.
Can I offer incentives for reviews to improve AI visibility?
Not in return for positive reviews. The FTC's 2024 rule bans incentives conditioned on a review's sentiment and bans fake reviews, and the UK has prohibited fake and concealed incentivised reviews since April 2025. Ask every buyer for an honest review, disclose any incentive clearly, and never suppress negative reviews.
How do I measure AI shopping visibility?
Track a fixed set of mostly unbranded shopping prompts by category, run each several times on each engine, and calculate the share of runs where your products are shown and recommended. Record price accuracy, the domain linked and cited sources too. Alongside that, segment assistant referral traffic and use Search Console's Generative AI performance report.
Is Winning the AI Shopping Shelf free?
Yes. Give a first name and a work email and the download starts at once. We also email you the link, and one person from the Noma team follows up a day later; reply and say so if you would rather not hear from us.
Why do you ask for a work email?
The playbooks are written for teams at companies, and a work address tells us which company is reading, so a follow-up can be about your brand rather than generic. Personal and temporary inboxes are not accepted.
What format is it?
A PDF of 51 A4 pages, laid out to print, with worksheets you can fill in by hand or copy into a document of your own.

More playbooks

Updated 20 October 2026 · Written by the Noma team · All resources