How to Get Your Products Recommended by ChatGPT, Gemini and Perplexity This Black Friday (2026 Guide)

Picture a shopper on a Tuesday in November. She types, “Best waterproof hiking boots under $150 for wide feet.” She does not scroll ten blue links. Instead, she reads one answer with four products, and she clicks the one that looks right. If your boots are not in that answer, she never sees them.

That shift is now measurable. Bain’s 2026 holiday survey of 1,105 US consumers, taken in August, found that 24% plan to start their holiday shopping on gen AI platforms such as ChatGPT, Claude and Gemini. That was 17% a year earlier. Meanwhile, Adobe forecasts that AI traffic to US retail sites will rise 130% year over year this holiday season. Both numbers matter. However, neither means AI is your main channel yet. Adobe itself says AI traffic is still modest next to other big channels. The reason to act is quality, not volume. In Adobe’s 2025 holiday data, visits from AI referrals converted 31% better than other traffic.

So what is the short answer? AI engines recommend products they can read, trust and sell. They read your product data, they check it against your page, and they weigh price, stock, quality and reviews. You control all of those inputs, and you can fix most of them this month.

One 2026 change makes this more urgent. OpenAI retired Instant Checkout in March, so ChatGPT now sends shoppers to merchant sites to buy. Walmart told WIRED that purchases inside ChatGPT converted three times worse than click-outs to its own site. As a result, your product page and your feed now do the selling. Here is what you will learn:

• The facts: what Adobe, Bain and OpenAI actually say, with their limits.
• The rules: how ChatGPT, Gemini and Perplexity each pick products.
• The plan: a dated, 29-day schedule to Black Friday.
• The fix list: feed fields, schema, reviews and crawler access.
• The test: a 12-prompt check you can run in one afternoon.

I also answer the questions store owners ask me most. Can I pay to be included? Is 29 days enough? Do I need Shopify? What if I sell on Amazon, eBay or Etsy? Keep reading, because the biggest mistake on this list costs sellers more than any missing tag, and it sits near the end.

What the data says about AI shopping this Black Friday

Start with the numbers, because the headlines blur three different measures: traffic growth, shopper intent and conversion. Here is each one, with its source and its limit.

Look at the tallest bar. Adobe expects the biggest jump on Thanksgiving Day, November 26, the day before Black Friday. So treat November 25 as your real feed deadline, not November 27. Adobe also expects $12.9 billion in US online sales on Black Friday and $15.1 billion on Cyber Monday, so the sales are real even if AI’s share is still small.

Read these together. Shoppers are moving toward AI for discovery, and the traffic converts well. At the same time, most of them still finish on a retailer’s own site. That is why Bain’s 13% of shoppers who plan to use a retailer’s own AI agent also matters. It tells you the buying step is staying close to the merchant. For wider context, see my AI search statistics for 2026.

How an AI shopping answer is built, according to OpenAI’s own help page

Most guides guess at this part. So I read OpenAI’s Help Center page on shopping in ChatGPT on October 29, 2026. It is the closest thing to a public rulebook. Below, each rule sits next to what it means for your store.

Notice what is missing. The page names no secret tag, no trick and no paid slot. It names data. That is good news, because data is the part you control.

How ChatGPT, Gemini and Perplexity choose products

Each engine has a different front door. Learn all three, then start with the one your customers use most.

ChatGPT: feeds drive discovery, your site closes the sale

OpenAI’s product feed reference lists nine required fields. They are an item ID, title, description, product URL, brand, seller name, image URL, availability and price. Optional fields include GTIN, MPN, sale price, shipping, return policy, star rating and review count. The feed currently targets the US. Also, OpenAI states that eligibility does not guarantee display.

The Instant Checkout story matters here. OpenAI launched in-chat buying in September 2025 and pulled back in March 2026. It said it would let merchants use their own checkout while it focused on product discovery. Walmart’s data shows the problem. It offered about 200,000 products in the feature, and its executive Daniel Danker told WIRED that in-chat purchases converted at roughly one-third the rate of click-outs. So treat ChatGPT as a discovery channel. Your job is to be in the shortlist and to win the click.

Gemini and AI Mode: your Merchant Center feed is the base

Google’s own developer page for its Universal Commerce Protocol tells merchants they can “use your existing Merchant Center account shopping feeds” to reach high-intent shoppers, and it says you remain the merchant of record. Checkout inside AI Mode and Gemini is still limited, so do not plan around it. Instead, make sure your Merchant Center feed and your Product markup are complete and agree with each other. Google’s documentation requires a name, an image and an offer with a price above zero and a currency. It recommends availability, GTIN, brand, reviews, shipping details and a return policy.

Perplexity: join the program, then earn the answer

Perplexity’s shopping runs on merchant data and on integrations such as Shopify and PayPal. Its merchant program works like a catalog submission, with details such as pricing, reviews, specs and images. Sources disagree on approval, so apply early and check Perplexity’s current sign-up page. In all three engines, one rule holds: complete, accurate, fresh data beats clever copy.

Shopify stores: what is automatic, and what is still yours

Shopify says its Agentic Storefronts feature puts eligible stores’ products into ChatGPT, Gemini and other AI chats, and that it is switched on automatically for eligible stores. It also says Shopify Catalog structures your product data so AI can read it, that AI orders land in your Shopify admin, and that you can see which sales came from AI. OpenAI’s help page backs up the ChatGPT part: no extra work is required from Shopify merchants.

Two cautions apply. First, “automatic” does not mean “complete.” A catalog can only structure what you give it, so fill in brand, GTIN, specs, shipping and returns. That is my advice, not a Shopify rule. Second, Shopify’s page promises buying inside the chat, while OpenAI retired Instant Checkout in March. Check which AI channels and checkout options your own admin shows before you plan around either.

Other assistants sell too. Retail TouchPoints reported on January 9, 2026 that Etsy and Shopify are among the early partners for Microsoft’s Copilot Checkout. An opinion piece in The Drum argues that Amazon’s Rufus is already racking up sales. I could not open either report in full, so I give no figures for them.

A 29-day plan from October 29 to Black Friday

Time is short, so each week has one job. Do the tasks in order, because later steps depend on earlier ones. Note that OpenAI says merchant updates can lag, so earlier is better.

The fix list, in order of impact

1. Make the product feed complete and true

Start with the nine required fields, then add the optional ones that help a model compare you. Put the right value in each field, and keep it matched to your page.

2. Mirror the feed in Product schema

Product markup lets Google and other crawlers read the same facts from the page. Google requires a name, an image and an offer with a price above zero and a three-letter currency. It recommends availability, GTIN, brand, review ratings, shipping details and a return policy. Below is a minimal example. Replace every value with your own, and keep it identical to the visible page.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Trailhead Waterproof Hiking Boot, Wide Fit",
  "image": "https://example.com/images/trailhead-wide.jpg",
  "brand": { "@type": "Brand", "name": "Trailhead" },
  "gtin13": "0123456789012",
  "offers": {
    "@type": "Offer",
    "price": "129.00",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock",
    "url": "https://example.com/products/trailhead-wide"
  }
}
</script>

Schema helps machines read your page. It does not guarantee a recommendation, and Google says no special markup is needed for its AI features. For the full picture, read my guide to schema markup for AEO and GEO.

3. Write product pages that an AI can quote

Models lift short, specific passages. So write them. Open with one sentence that says what the product is, who it suits and its price. Next, add a spec table, then a short “who it is not for” line. After that, answer five real buyer questions, such as sizing, materials and returns. Finally, keep the price in the page HTML, not only in a script. My guide to how to get cited by ChatGPT shows the pattern in detail.

4. Build proof you can show

“Quality” is one of OpenAI’s stated ranking factors, and reviews are the easiest proof to supply. Ask every recent buyer for a review, and make it simple. Never buy or invent reviews. They break platform rules and can get your feed removed. Beyond reviews, earn a few honest mentions in places buyers read, such as gift guides, comparison articles and community threads. I explain why third-party mentions matter in why Google rankings do not match ChatGPT mentions.

5. Let the right bots in

Blocking the wrong crawler can remove you from AI answers. OpenAI documents three bots with separate jobs. OAI-SearchBot surfaces sites in ChatGPT search, and sites that opt out will not appear in ChatGPT search answers. GPTBot is for training and can be blocked without affecting search. ChatGPT-User fetches pages when a person asks. Perplexity separates PerplexityBot for search from Perplexity-User for user requests. Check your robots.txt, then check your firewall and bot-protection settings, because many stores block these crawlers by accident. My AI crawlers and robots.txt guide has the full list.

6. Get your Black Friday offer data right

Black Friday breaks feeds. Prices change, stock sells out and codes expire. So prepare three things now. First, load your sale prices into the feed with a clear sale_price that is lower than price. Second, plan who updates stock during Thanksgiving week. Third, keep your page price, feed price and schema price identical. A mismatch can cost you the recommendation.

Your AI shopping readiness scorecard

Score each line 0 (no), 1 (partly) or 2 (yes). Be honest, because the lowest scores are your first tasks.

If you score 0 to 8, fix the feed and crawler access first. If you score 9 to 15, you are close, so focus on reviews and page copy. If you score 16 to 20, spend your time on testing and Black Friday updates.

Short on time? What can realistically be done before November 27

I do this work for clients, so I will be direct about timing. The table uses the prices and delivery times on my pricing page, checked October 29, 2026. Some options will not fit before Black Friday, and I say so.

None of these buys a placement, and I cannot promise one. They fix the inputs the engines say they read. If you are unsure where to start, run the free audit first. It costs nothing and shows which row above you actually need.

If you’re starting from zero: no store, no stock, no sales yet

Everything above assumes you already sell something. AI can only recommend what exists. It needs a product, a price, stock and proof. So if you have none of those yet, do not wait for the perfect store. Start small and sell your first items now.

That is the gap my own guide fills. It is called How to Make Your First Online Sales Before Black Friday (Even If You’re Starting From Zero). It is a 14-day plan, with one clear task per day, for people in the US and Canada who want their first listings live this week. It is a product I made, and I sell it on Gumroad.

It also ties into this article. Module 4 teaches the same habit AI shopping rewards: clear titles, honest descriptions and facts you have checked. Marketplaces work like search engines, and so do AI assistants. Complete, accurate listings help in both.

What if you sell on Amazon, eBay, Etsy or Walmart?

Marketplace sellers control less of the data. However, you still control the parts models read. Write exact titles with brand, type, model, size and color. Fill every attribute field. Use sharp photos, keep stock accurate and ask for reviews. Also, if you have your own site, build it in parallel, because Bain found 60% of shoppers plan to start on retailer or brand websites. I could not verify which marketplaces feed which AI engines today, so do not assume. Test your own listings with the prompts below.

The 12-prompt test: see where AI puts you today

You can measure this yourself. Choose four product searches that a buyer would type, then run three styles of each on ChatGPT, Gemini and Perplexity.

Run each prompt three times, because AI answers change. SparkToro and Gumshoe.ai found the same brand list in under 1% of repeated runs, so count how often you appear. Then check what you see against your feed. If a price or fact is wrong, fix the source. My guide to checking what ChatGPT says about your business gives the full method, and the comparison of AI visibility tools helps if you want it automated.

Next, track the visits. In Google Analytics 4, look at sessions by source for referrers such as chatgpt.com, perplexity.ai and gemini.google.com. Compare their conversion rate with your other traffic. That is how you will know whether this work paid off.

7 mistakes that keep products off AI shortlists

1. Price and stock mismatches. The feed says one thing, the page says another.
2. Blocked crawlers. A firewall or robots.txt rule hides you from search bots.
3. Thin product data. No GTIN, no brand, no specs and a weak description.
4. Fake or seeded reviews. They risk removal and damage trust.
5. Waiting for in-chat checkout. It was retired. Build for click-through instead.
6. Launching late. OpenAI notes updates can lag, so a Thanksgiving-week change may miss the rush.
7. Treating AI as a trick. The engines publish their rules, and the rules reward real data. The biggest mistake is not a missing tag. It is leaving product data to chance until the week of the sale.

What to do next

Do the first two tasks today. Run the 12-prompt test, then fix your crawler access. Those two steps cost nothing and show you where you stand. If you want a second pair of eyes, I run this check as part of a free AI visibility audit. See the pricing page, my pages on AEO and GEO, or tell me what you want to be found for. For e-commerce proof, read how Lens & Co reached first place in ChatGPT answers and how LAGE Lightings moved from seventh to first.

And if you are still at the starting line, the 14-day plan is here: How to Make Your First Online Sales Before Black Friday.

How I checked, and the limits

I read the primary sources where they were public: Bain’s holiday outlook page, OpenAI’s product feed reference and bot documentation, Google’s structured data and UCP pages, and Perplexity’s bot documentation. Adobe’s forecast is cited through Digital Commerce 360, and Adobe’s 2025 figures through Adobe’s own blog. I read OpenAI’s Help Center page on shopping in ChatGPT, the Digital Commerce 360 report on Adobe’s forecast and Shopify’s Agentic Storefronts page in a normal browser session on October 29, 2026, and captured the screenshots myself. Bain’s page now serves a bot check, which I did not bypass, so I could not capture it. I also could not open Forbes, WIRED, Perplexity’s merchant page or the full reports on Copilot Checkout and Rufus, so those details come from reports and are marked as such. Surveys measure intent, forecasts can miss, and platform rules change often. Nobody can promise an AI recommendation, so treat this as a plan to improve your odds.

Related reading: SEO, AEO and GEO explained and AI citation versus AI recommendation.