AI Citations by Subvertical: The Aleyda Solis Study That Reveals Patterns Beyond Product Pages
Summarize this article with AI
Why your product pages, however perfect, stay invisible to AI
I review 15 sites a week.
They all have the same problem.
Polished product pages. Flawless schema markup. Well-structured PLPs. Yet when I test with AI Overview or SearchGPT, they’re invisible. Not cited. Not referenced. Why?
Because they haven’t understood what AI is really looking for.
We measure this frustration every day in audits. A cosmetics site with 800 references, perfectly indexed. Zero citations on AI queries. Meanwhile, AI responses showcase influential blogs, tutorial videos, forum comparisons. Not a single page from the site.
This is where Aleyda Solis’s study, published May 12, 2026, brings crucial insight. She analyzed AI citation sources from 25 leading sites across 5 e-commerce subverticals in the United States: général marketplaces, beauty, fashion, consumer electronics, sports. The finding is clear: the majority of cited pages are not product or category pages. They’re guides, support pages, return policies, sizing pages, comparisons.
In other words, AI cites what reassures the buyer, not necessarily what triggers the direct purchase.
This reversal changes the game. You optimized your PDPs for classic Google. But AI is looking for something else. The good news? It’s a massive opportunity for e-commerce businesses willing to break free from the « product page = SEO » logic.
The 6 patterns redefining AI SEO in e-commerce
The study identified 6 recurring patterns. Not hypotheses—concrete observations from Semrush Enterprise citation data.
- Pattern 1: AI citations aren’t limited to product/category pages. Guides, support content, educational material, and local pages appear massively. 47% of citations identified in the study are not transactional pages.
- Pattern 2: A layer of shared citation sources exists, but its role shifts by vertical. For example, forums are heavily cited in electronics, sizing guides in fashion.
- Pattern 3: The mix of sources evolves based on the proof being sought. AI cites return policies when a user asks « is this site trustworthy? », and tutorials for « how do I choose ». We observe that 60% of responses include at least one non-commercial page.
- Pattern 4: Each subvertical has a different buyer uncertainty pattern. In beauty, uncertainty centers on compatibility (skin type); in electronics, on technical specs; in fashion, on sizing. Understanding this pattern tells you what content to create.
- Pattern 5: Général marketplaces are the only vertical where competitors cite each other. In other catégories, external citations come from media, influencers, communities. A strong signal: your authority won’t come from peers, but from recognized third parties.
- Pattern 6: Even category leaders have a minority citation share on queries about their own products. AI doesn’t favor the brand selling—it seeks the source answering best. This completely overturns the domain authority concept.
These patterns are verifiable in Aleyda Solis’s original study. The point isn’t mastering them all in theory, but extracting concrete actions by sector. That’s what we’re doing next.
What each subvertical must optimize as a priority
Patterns become actionable when tied to a specific vertical. Here’s what Aleyda Solis’s analysis teaches us, sector by sector.
Général marketplaces
AI cites policy pages heavily, customer service, fee comparisons, and trust guides. One marketplace client saw AI citations jump +340% after publishing an ultra-detailed refund policy and a « How to Spot a Trustworthy Seller » page.
Beauty & skincare
Routines, skin-type compatibility, before-and-afters, video tutorials. These are the most-cited pages. A cosmetics site gained 1,200 AI clicks/month by creating 12 routine guides and 5 ingredient comparisons, without touching its product pages.
Fashion & apparel
Sizing guides, lookbooks, fit advice, return and exchange policies dominate. A clothing brand cut its return rate by 22% after building a multimodal sizing guide (video + chart + chatbot). AI citations skyrocketed.
Consumer electronics
Advanced spec sheets, device compatibility, repairability, community forums. AI almost never cites a standard product page. It prefers technical comparisons or user tests.
Sports & outdoors
Usage guides, performance comparisons, expert reviews, maintenance tips. An outdoor gear site tripled its AI traffic after publishing 8 category selection guides by activity.
Pages to create or strengthen right now
Let’s move past theory. Here’s a checklist you can act on immediately, ranked by likely impact on AI citations.
- 1. Selection guides – for each major category, build a page helping customers choose between 3 or 4 products by objective criteria (skin type, body shape, use case). AI loves structured comparisons.
- 2. Policy pages – returns, refunds, warranty, shipping. Make them readable in 3 seconds, with visual hierarchy (tables, icons). These are trust proofs cited heavily.
- 3. Sizing/compatibility guides – embed dynamic tables, videos, FAQs, product links. A well-built sizing guide can become your site’s most-cited page.
- 4. Educational and tutorial content – beauty routines, electronics setup tutorials, sports gear maintenance tips. Show usage, not just products. This is what reassures buyers.
- 5. Detailed customer service and FAQ pages – don’t just offer a form. Build structured answers organized by topic, linked to your policies. AI will pull from these.
- 6. Community spaces (forums, enriched reviews) – in electronics and sports, external citations often come from forums. Provide a controlled space for customers to share expériences.
The classic mistake: creating all these pages with no connections between them. Solid semantic linking is essential. Each guide page must point to products, policies, comparisons. Each product page must link back to its guides. It’s a trust cocoon that AI will read as a coherence signal.
Where to start? Audit your current AI citations on 50 key queries. Note which pages are cited. Are they yours? If not, what type of page is missing? That’s your first project.
Real case: 1,200 monthly clicks gained without touching product pages
A French cosmetics site: 800 SKUs, 40,000 organic sessions/month. Well-written product pages, flawless schema, HD photos. Zero classic SEO problems.
Yet on 50 AI queries tested (type « best serum for oily skin », « anti-wrinkle routine »), this site had zero citations. AI cited beauty bloggers, comparison sites, forums. Not a single page from the site.
Quick analysis: the site had no guided content. No routines, no comparisons, no skin-type pages. It sold creams, not solutions.
We created:
- 12 routine guides by skin type (oily, dry, sensitive, mature), each linked to matching products.
- 5 ingredient comparisons (hyaluronic acid vs retinol, etc.) with simplified explanations.
- 3 dermatologist-sourced advice pages.
- 1 cross-reference table « skin type x need ».
Total cost: 2 months editorial work, €6,000. Zero changes to product pages.
Results after 4 months: +1,200 AI clicks/month (SearchGPT, AI Overviews). +32% incremental revenue from these visits. Guide page conversion rate: 4.8%, vs 2.1% for category pages. Classic organic traffic also grew 17%, as these new pages captured informational queries the site was missing.
The breakthrough? The site stopped selling products. It sold clarity and confidence. Exactly what AI seeks.
The mistake I see everywhere and the question that changes everything
Many e-commerce businesses think optimizing product pages is enough. Product schema, optimized titles, rich descriptions. Necessary, not sufficient.
The mistake? Assuming AI makes purchases. AI doesn’t check out. It helps a human decide. It cites what reduces uncertainty. A sizing guide, a clear return policy, an objective comparison. Not a standard product page.
So the question that changes everything, when you work on AI SEO: « Which page would best answer my customer’s pre-purchase question? » Not « Which page do we want to sell? »
Start with that question, and your content strategy shifts. You’ll build pages AI cites naturally, because they’re the best available answer. And these pages will sell—not by push, but by the strength of good advice.
I’m not selling you a method. I’m showing you the pages.
Request an AI visibility audit
I review your site and show you exactly which pages are missing from AI citations. In 45 minutes, you’ll know what to do.
Book a strategic call — 45 minFrequently Asked Questions
What’s the difference between a product page and a guide page in AI citations?
A product page describes an item. A guide page helps you choose, use, compare. AI cites the guide because it answers the buyer’s question—the product page adds no decision-making value.
Should I delete my product pages?
Absolutely not. They remain essential for classic SEO and purchase conversion. The goal is complementing them with content pages that reassure buyers upfront.
Do backlinks still matter in AI citations?
Aleyda Solis’s study doesn’t focus on backlinks, but we observe that AI cites diverse sources, often media or communities. A healthy, diversified link profile can indirectly strengthen your guide authority.
How long until results appear in AI responses?
With a structured strategy, first citations can emerge in 4 to 6 weeks if pages are indexed and useful. Traffic impact typically shows in 3 to 4 months.
Does this approach work for classic Google SEO too?
Yes. Creating guides and support pages improves rankings on informational queries and drives qualified organic traffic, often with higher conversion rates than PLPs.
