AI Overviews Now Answer Product Queries — Here's How to Stay Visible

The AI Overview Shift That Changes Everything for Product Sellers

If you sell products online, the most important Google update of 2025 wasn’t an algorithm change — it was a behavioral shift in what triggers AI Overviews.

For months, Google’s AI-generated summaries were mostly an informational play. You’d see them for “how does photosynthesis work” or “what is machine learning.” They were interesting, but they didn’t threaten ecommerce traffic directly.

That changed fast.

According to Semrush Sensor data, the percentage of informational keywords triggering AI Overviews dropped from 89.03% in October 2024 to just 57.16% by October 2025. The gap — now roughly 43% of AI Overview triggers — is filled by navigational, commercial, and transactional queries. That means product comparisons, “best of” lists, pricing lookups, and buying guides.

In plain terms: Google’s AI is now answering the exact queries your customers used to click through to your product pages for.

What the Data Actually Shows

Here’s what the research from Semrush and Pew Research Center tells us about the current state of AI search:

• AI Overviews now appear for approximately 12.95% of search queries in the U.S. market on average, according to Semrush Sensor’s ongoing analysis of millions of keywords.
• Users click traditional organic results just 8% of the time when an AI Overview is present, compared to 15% when there’s no AI Overview — a finding from Pew Research Center.
• Users click links within the AI Overview itself only about 1% of the time. That means even when your brand is cited in the AI answer, the direct click-through is minimal.
• AI Overviews use a technique called query fan-out, where Google’s system runs multiple related searches simultaneously to synthesize a comprehensive answer. A single query like “best running shoes for flat feet” might trigger sub-queries about arch support, durability, price comparisons, and podiatrist recommendations.

Google itself states that its recommendation for optimizing for AI features is to “apply the same foundational SEO best practices for AI features as you do for Google Search overall,” as documented in their Search Central documentation.

But “foundational SEO” means something very different when the goal is being cited by an AI rather than being clicked by a human.

Why This Is Different From Traditional SEO

The old SEO contract was simple: rank high, get clicks, convert visitors. AI Overviews break that contract in three ways:

1. Zero-click answers eliminate the middleman. When Google’s AI says “The best CRM for small businesses is HubSpot, followed by Zoho and Salesforce,” the user often gets what they need without clicking anything. Your product page doesn’t even get a chance to make its case.

2. Citations replace rankings. In AI Overviews, being cited as a source matters more than your position on the page. The AI pulls from multiple sources, synthesizes them, and presents a unified answer. Your goal is to be one of those sources — ideally the primary one.

3. Query fan-out multiplies the surface area. Because Google’s AI runs multiple sub-queries for complex questions, your content needs to address not just the main query but the related sub-queries too. A product page optimized for “ergonomic office chair” also needs to cover posture benefits, lumbar support comparison, price-to-value analysis, and assembly complexity — because the AI might check all of those.

The Structured Data Advantage

This is where structured data becomes your single most important technical asset for AI visibility. Google’s Product structured data documentation specifies the properties that help Google understand your products — and those same properties feed the AI models that generate AI Overviews.

Here’s what a citation-ready Product schema looks like:

For a citation-ready Product schema, make sure each product page exposes these fields in clean JSON-LD:

• Product identity: name, description, image, brand, GTIN, and MPN where available.
• Offer data: URL, price, currency, availability, and seller details.
• Trust signals: aggregateRating with both ratingValue and reviewCount, plus review data when it is accurate and eligible.
• Entity consistency: the same brand, product identifiers, and offer details that appear visibly on the page.

These fields give Google’s AI machine-readable signals about what your product is, who makes it, what it costs, and whether buyers trust it. When the AI is deciding which products to include in its generated answer, these signals act as confidence markers.

Google’s structured data documentation recommends including all required and recommended properties for the Product type to maximize eligibility for search features — and by extension, AI Overviews.

Five Action Steps for Product Sellers

1. Audit Your Product Schema for Completeness

Most ecommerce product pages have basic structured data — a product name, price, and maybe an image. But the AI models need more context to cite you with confidence.

Check every product page for:

• Brand property with proper Brand type nesting
• GTIN and MPN identifiers where available
• AggregateRating with both ratingValue and reviewCount
• Offer details including availability status
• Description that’s factually rich, not marketing fluff

Use Google’s Rich Results Test to validate your schema, then cross-reference against the Product type specification on Schema.org.

2. Write Citation-Ready Product Descriptions

AI Overviews don’t pull marketing copy. They pull factual, specific statements that answer the sub-queries generated by query fan-out.

For each product, make sure your description addresses:

• What it is — category, materials, key specifications
• Who it’s for — use cases, target audience, common scenarios
• How it compares — positioning relative to alternatives (without naming competitors directly)
• What reviewers say — aggregate sentiment and specific praise points
• What it costs and why — value proposition tied to features

Structure these as self-contained paragraphs that an AI could extract independently. Each paragraph should make sense on its own, because that’s exactly how AI Overviews consume and republish your content.

3. Build Third-Party Brand Mentions

The original GEO research paper from Princeton University demonstrated that Generative Engine Optimization techniques can boost content visibility by up to 40% in AI-generated responses. One of the strongest signals? Mentions of your brand on third-party websites.

AI models learn about entities — products, brands, people — from the broader web, not just your own site. If your product is mentioned in buying guides, review roundups, blog posts, and forum discussions across multiple authoritative domains, the AI is more likely to recognize your brand as a legitimate entity worth citing.

Tactical moves:

• Get your products listed in independent review roundups and comparison articles
• Respond to media requests on platforms like HARO and Featured to earn expert mentions
• Encourage customers to leave reviews on third-party platforms (not just your own site)
• Create shareable data, original research, or unique angles that give journalists a reason to reference you

4. Optimize for the Fan-Out, Not Just the Query

Because AI Overviews use query fan-out to run multiple sub-queries, your content strategy needs to cover the constellation of related questions — not just the primary keyword.

If you sell “organic dog treats,” the AI might fan out to:

• Are organic dog treats healthier than conventional ones?
• What ingredients should I avoid in dog treats?
• What’s the price difference between organic and standard dog treats?
• Which brands do veterinarians recommend?

Create content that addresses each of these sub-queries — either on your product pages (in FAQ sections) or on supporting blog posts that link back to your products. The more of the fan-out you cover, the more likely the AI is to cite you as a comprehensive source.

FAQPage schema is particularly effective here. It gives you a structured way to present question-answer pairs that AI Overviews can extract directly.

5. Track Your AI Visibility Separately From Traditional SEO

Traditional SEO tools track rankings and clicks. AI visibility requires different metrics:

• Citation rate: How often does your brand or product appear in AI-generated answers for relevant queries?
• Share of AI voice: When the AI recommends products in your category, what percentage of mentions do you capture versus competitors?
• Sentiment of citation: When cited, does the AI frame you positively, neutrally, or negatively?
• Query coverage: Across the full fan-out of related queries, how many do you appear in?

This is exactly the gap that FirstShelf.AI was built to address — giving product sellers the ability to monitor and improve their visibility specifically within generative AI platforms, not just traditional search.

The Bigger Picture: From Clicks to Citations

The shift from informational to commercial AI Overviews isn’t a temporary phase — it’s the trajectory. Google is investing heavily in its Shopping Graph and integrating product data with AI Overviews and AI Mode. The direction is clear: more product discovery will happen inside AI-generated answers, not only on traditional result pages.

The sellers who thrive in this environment won’t be the ones with the best backlink profiles or the most keyword-stuffed pages. They’ll be the ones whose products are the most legible to AI — structured, cited, mentioned, and recognized as entities across the web.

The data is clear: clicks are declining, but citations are growing. The question isn’t whether to optimize for AI-generated answers — it’s how fast you can make the shift.