By Tarun Babbar, Co-founder Gravton Labs | October 2026

Imagine someone asks ChatGPT: “What’s the best 65-inch LED TV under $1,000 for watching sports?”

The answer might compare TVs, mention reviews, recommend models, and suggest where to buy. One conversation now covers most of the shopping journey, from research to choosing a store.

For multi-brand marketplaces, getting your product catalog recommended by AI isn’t the same as getting your store recommended for the actual purchase.

AI could recommend a TV you sell, but cite another retailer for price or stock. It might use your buying guide but not name your store. Or it might recommend your marketplace for services like installation or delivery.

So, retailers compete on two things: what they sell (catalog visibility) and why someone should buy from them (retailer visibility).

At Gravton Labs, we looked at how AI makes these choices. We examined which sources get cited, when retailers are mentioned, and which content makes a difference. Here are five key areas for multi-brand retailers and marketplaces. (PS: the rest of the article may use multi-brand retailers and marketplaces interchangeably.)

What We Studied

We analyzed 380 AI shopping answers and 1,691 citations from three retail categories in India and the Middle East, mainly electronics. For electronics, we reviewed 311 answers and 1,417 citations about smartphones and TVs. We asked shopping questions in ChatGPT and Google AI Mode, checked which sources they cited, who owned them, and whether they named brands or retailers. We sorted pages by type, product listings, categories, services, guides, reviews, comparisons, offers, and videos. We also checked nearly 1,500 pages ChatGPT found but didn’t cite.

Our goal: identify which sources drive product recommendations, which drive retailer mentions, and what marketplaces can learn.

1. Product Discovery vs. Retailer Discovery: Two Different Challenges

When shoppers ask AI what to buy, the answers usually come from reviews, comparison sites, and buying guides. In our study, media and price-comparison sites made up 71% of editorial citations.

But retailer-owned content really matters when AI says where to buy. In electronics, over 90% of answers that named a retailer had at least one page from that retailer’s site backing it up, usually with product details, prices, or delivery info. For example, a review might help AI pick a TV, but a retailer’s product page often tells it where you can buy it.

Thus, multi-brand retailers and marketplaces should track the ‘product catalog being recommended’ and ‘the retailer being recommended’ separately.

2. Buying Guides Help With Research—But Product Pages Drive Retailer Mentions

We looked at shopping questions that didn’t name a retailer, and checked how often retailers were mentioned when AI cited different page types.

Retailers were mentioned in 88% of answers citing their product pages and 86% citing category pages, compared with 57% for offer pages and just 10% for buying guides, comparisons, or reviews.

Our takeaway: product and category pages were much more likely to get retailers mentioned than buying guides. That’s likely because product pages link specs and recommendations directly to prices and availability, while buyers research products around specifications.

A buying guide about OLED vs. Mini LED may help AI answer research questions, but it doesn’t always say why to buy from your store. AI can use the info and skip the retailer.

Buying guides are still valuable, but they serve a different purpose. They should connect more clearly to PDPs (Product Display Pages).

Keep making helpful, up-to-date buying guides, but link them directly to product listings, prices, and offers. This helps tie advice to buying choices and reflects a high-performing strategy.

3. Category Landing Pages Should Reflect How People Shop (Hint: Pre-Activated Filters)

Most marketplaces organize by category and brand. But AI shopping questions are more specific, like “best 75-inch TVs under $1,500” or “Mini LED TVs for bright rooms.”

These questions combine product features, budgets, and customer needs.

Category and listing pages made up about 15% of all citations in our study. When AI cited a retailer’s category page for questions without a store name, that retailer showed up in 86% of answers. Pages grouped by specific product features best matched shopper questions.

For example, a page with 75-inch Mini LED TVs under a certain price is more useful for a specific question than a generic TV category page.

Retailers should group products by customer needs, not just by catalog structure. For electronics, that could mean grouping by screen size, technology, price, gaming, or use case.

You don’t need thousands of new pages, start with key customer-intent groups. Make sure these pages have accurate products, good filters, and up-to-date info.

4. Services and Policies Can Set You Apart in AI Shopping

In a high-consideration category like electronics, support and policy pages accounted for about 18% of all citations, the most of any page type. With store info and services, that went up to 23%.

These pages included delivery, installation, financing, returns, warranties, and other buying services.

Think about a shopper who picked a TV but now wants wall-mount installation, payment plans, or fast delivery. The choice shifts from what to buy to where to buy.

Our research shows service and policy pages drive these decisions. Clear info on installation, financing, delivery, and returns helps AI link benefits to your store.

But retailers often split this info across FAQs, banners, policies, and checkout, making it hard to find.

Make it easy to find and understand your services, like installation, financing, exchanges, warranties, and local delivery. Provide clear info on prices, eligibility, restrictions, and service availability.

5. Some Content Gets Missed If AI Can’t Access It

We also found technical issues can block AI from seeing your content.

On one electronics site, buying guides used JavaScript to load the main text, so the HTML had no content and the sitemap missed the blog posts. Another site’s new guides weren’t in its sitemap at all.

None of these guides showed up in ChatGPT results. For both retailers, less than 1% of their own site citations were for guides, comparisons, reviews, or best-of lists.

We can’t say tech issues are the only reason, since AI systems vary and sitemaps don’t guarantee discovery. But retailers should check for tech barriers before making more content.

Audit your most valuable content for accessibility before making new guides. Check whether key info is in the HTML, whether pages are discoverable via links and sitemaps, and whether product details are easy to find without extra clicks or scripts.

Then see which pages AI actually finds, cites, or misses. Fixing existing access issues can be more valuable than posting lots of new content.

What This Means for Multi-Brand Marketplaces

Improving AI visibility for marketplaces requires a different approach than for product brands. For a product brand, being discovered or being shortlisted is already valuable. However, marketplaces must not only get those products recommended but also establish themselves as a preferred place to buy, often competing with several retailers selling the same products.

Our research shows that different sources play distinct roles. Independent publishers (including those supported by brands themselves) shape product research, while retailer-owned pages provide evidence to support retailer recommendations. Product information, pricing, availability, and purchasing services all contribute to this process.

For marketplace operators, I would prioritize three things: measure product and retailer visibility separately, connect product information with purchasing options, and address technical issues limiting content discovery.

An equally important area is catalog enrichment: maintaining accurate, complete, structured, and accessible product information for AI assistants through PDPs and catalog feeds. This deserves a deeper examination and is outside the scope of this study.

About the Research

Gravton Labs is a Toronto-based AI visibility platform for scaled e-commerce brands, multi-brand retailers, and marketplaces. This study was done in October 2026, using shopping questions on ChatGPT and Google AI Mode in India and the Middle East. The findings describe patterns in the answers and citations we studied. They don’t prove cause and effect or show all AI shopping activity.

Originally published on LinkedIn by Tarun Babbar, Co-founder of Gravton Labs.