How to Get Your Products Recommended by ChatGPT

CMSmart (cmsmart.net, part of Netbase JSC) sells a WooCommerce product designer plugin for print shops. In Leadee's tracker run of 2026-10-25, OpenAI's gpt-5 with web search, called through the official API, named CMSmart in 2 of 3 runs for "best WooCommerce product designer plugin for print on demand" and in the only run for "Fancy Product Designer alternatives for WooCommerce". That is 3 of 4 runs. The OpenAI account ran out of credit after 25 runs across all brands, so treat this as an indicator, not a rate. The same day, across all US questions, Perplexity named CMSmart in 13 of 24 runs and Google AI Mode in 8 of 24 (95% Wilson intervals are wide at this size). The full numbers are in the CMSmart case study.

On 2026-11-09 we read the product pages behind that product, plus two Netbase-owned homepages (part of Netbase JSC) for comparison. The CMSmart plugin pages declare SoftwareApplication markup with a USD 290 offer, and one also declares rating markup (4.9 from 223 reviews; we did not audit the reviews behind it). The "Fancy Product Designer alternative" article is marked up as an article with no product. The Printcart and Web2Print homepages, also Netbase-owned (part of Netbase JSC), have no product markup at all.

This does not show that the markup caused the mentions. CMSmart also has years of content, many citations from other sites, and the most-cited domain in its niche. We only know what was on the page and what the engines said. Passage quality was ordinary everywhere: average Citation Scores of 0.25-0.26, with few passages carrying a number.

Product recommendations come from questions such as "best X for Y", "X alternatives" and "X vs Y". Write 10-20 of them in your buyers' words, for the use cases you actually win. Our tracker re-count for Netbase JSC brands found that unbranded questions are much harder to appear in than questions that name the brand. Finding the prompts customers ask ChatGPT shows how to build the list; the free ChatGPT Prompt Finder suggests starting points.

Ask each question several times and note which products are named, in what order, and which pages are cited. One answer is one sample; the same question can name different products in the next run. Checking if ChatGPT mentions your brand covers the method.

The cited pages are your to-do list. In our 2026-10-25 run for CMSmart's niche, the most-cited domains (pooled across Perplexity, Google AI Mode and OpenAI) were cmsmart.net, woocommerce.com, fancyproductdesigner.com, zakeke.com and wordpress.com. In other words, the brand's own site, the platform's marketplace and market players' sites.

An AI answer that recommends products needs facts to compare: what the product does, who it is for, price, platform or compatibility, key limits, support. Put them on the page as text, not only in images, sliders or a PDF. A crawler reads the HTML; it does not open your product video. Run the page through the ChatGPT Visibility Checker to see whether OpenAI's bots can reach it and what they find.

Write each key fact as a sentence that stands on its own: "The plugin costs USD 290 and works with WooCommerce" rather than "Get it now for just $290!". Our guide on writing pages ChatGPT can quote shows how, and the passages on our test pages show the gap: few carried a number.

Mark the page up with schema.org Product (for physical goods) or SoftwareApplication (for software), with name, description, brand, an Offer with price and currency, and availability. Only add rating markup if the ratings come from real reviews shown on that page. Every value in the markup must match what a visitor can see.

Build it with the ChatGPT Schema Generator and check it with the ChatGPT Schema Validator. The schema markup guide explains the Organization markup that ties your products to your company. OpenAI does not publish how ChatGPT uses page markup, so treat markup as clear labelling, not a ranking switch.

Questions like "Fancy Product Designer alternatives for WooCommerce" are where CMSmart was named in our run, and CMSmart has a page that answers exactly that question. Write comparison pages for the market players buyers compare you with. Be fair: say where the other product is the better fit. A fair comparison gives buyers, and any answer engine reading the page, usable facts about both products; a page that only praises the seller gives them little to compare.

Make sure your product has a complete, current listing on the marketplaces and review sites that are cited for your questions in Step 2, such as the app store of the platform you sell on. Use the same product name, price and description everywhere. Ask real customers for reviews; never buy or write them.

OpenAI has a product feed program for merchants. Its commerce documentation (checked 2026-11-09) says onboarding product feeds in ChatGPT is currently available to approved partners, who apply through a form at chatgpt.com/merchants. Feeds use a documented schema, are uploaded over SFTP as JSONL or CSV/TSV files, and should include titles, descriptions, images, price and availability. OpenAI recommends a full snapshot at least once a day.

The documentation we read does not say how feed products are ranked against products ChatGPT finds on the open web, or whether a feed is needed to appear in its answers. Do not skip Steps 3-6 because you applied for a feed. Also check that OAI-SearchBot, which OpenAI says surfaces websites in ChatGPT's search features, is allowed to reach your product pages.

Re-run the questions from Step 1 every week or two, the same number of times, and count how often your product is named and your page cited. Expect slow change and noisy results. With a handful of runs, one extra mention can look like a big jump. Leadee's tracker does this through OpenAI's official API and reports each rate with a confidence interval; the ChatGPT SEO overview shows the rest of the work.