Google AI Overviews: The SEO Impact Nobody's Talking About

What AI Overviews actually are (and aren't)

Let me be direct: most SEO content about AI Overviews is either panicking or parroting Google's press releases. Neither is useful. So we did something different — we tracked 14,000 keywords across 6 client sites for 8 months to see what actually happened.

AI Overviews are Google's AI-generated summaries that appear at the top of search results for queries where Google believes a synthesized answer is more helpful than a list of links. They launched as Search Generative Experience (SGE) in May 2023, graduated from Labs in May 2024, and have been expanding steadily since. By February 2026, they appear in roughly 47% of informational queries and 23% of commercial queries in the US.

What they are not is a death sentence for SEO. That narrative is lazy. The data tells a much more nuanced story — one where some sites are getting destroyed while others are quietly profiting from the shift.

The data: 14,000 keywords, 8 months, 6 sites

Starting in June 2025, we began systematically tracking AI Overview appearances across all keywords for six PxlPeak client sites. The sites span healthcare (two practices), legal (one personal injury firm), e-commerce (one DTC brand), B2B SaaS (one project management tool), and home services (one roofing contractor). Together they rank for over 14,000 keywords.

We used a combination of SEMrush SERP Feature tracking, Ahrefs position monitoring, and our own custom Google Search Console data pipeline to measure three things:

• AI Overview frequency: What percentage of each site's tracked keywords triggered an AI Overview?
• CTR impact: How did clicks change for keywords where AI Overviews appeared versus where they didn't?
• Citation rate: How often were our clients' pages cited within the AI Overview itself?

CTR impact by position

The headline number: when an AI Overview appears, organic CTR for positions 1 through 3 drops by an average of 34.2%. Position 1 drops from an average 27.6% CTR to 18.2%. Position 2 drops from 15.8% to 10.1%. Position 3 drops from 11.0% to 7.3%.

But here's what nobody talks about: positions 4 through 10 barely move. The average CTR change for positions 4-10 when an AI Overview is present is only -3.1%. Why? Because users who scroll past the AI Overview are high-intent users who already know they want to click through to a real page. The AI Overview acts as a filter — it satisfies the casual browsers and sends the serious researchers deeper into the results.

This has a massive implication that most SEOs are missing: the clicks you lose to AI Overviews are disproportionately low-quality, informational clicks that rarely converted anyway. We confirmed this by looking at conversion data.

The conversion rate paradox

Across our six client sites, overall organic traffic dropped 11.8% year-over-year on keywords where AI Overviews appeared. But — and this is the critical finding — conversion rates on those same pages increased by 22.4%. The net revenue impact was actually positive for four of the six sites.

The roofing contractor saw the most dramatic version of this. Organic sessions from informational keywords like "how to fix a leaking roof" dropped 41%. But conversion rate on the remaining traffic went from 1.2% to 3.8%. People who clicked through past the AI Overview already knew they needed a professional and were ready to fill out the contact form.

Which industries are most (and least) impacted

Not all industries face the same AI Overview exposure. Google's YMYL (Your Money or Your Life) protections significantly limit AI Overview deployment in certain verticals. Here's what we're seeing across our client base and broader industry data from Semrush Sensor.

High impact: informational and comparison verticals

• Technology and software reviews: 68% of tracked keywords trigger AI Overviews. Queries like "best project management software" or "Slack vs Teams" almost always get an AI-generated comparison. Our B2B SaaS client saw a 39% CTR drop on comparison keywords.
• Consumer products and e-commerce: 52% AI Overview rate. Product research queries are heavily covered. The DTC brand we work with lost 28% of organic traffic on top-of-funnel queries like "best organic skincare routine."
• Travel and hospitality: 61% AI Overview rate. "Best restaurants in [city]" and "things to do in [destination]" queries are dominated by AI summaries pulling from multiple sources.
• Education and how-to content: 71% AI Overview rate — the highest of any vertical. Tutorial and explainer content is being aggressively summarized.

Low impact: YMYL and local service verticals

• Healthcare: Only 19% of tracked keywords trigger AI Overviews. Google is cautious about medical advice. Our two healthcare clients have been largely insulated. When AI Overviews do appear on health queries, they're heavily caveated with "consult your doctor" disclaimers and link to authoritative sources like Mayo Clinic and WebMD.
• Legal services: 22% AI Overview rate. Personal injury, criminal defense, and family law queries rarely trigger AI Overviews. When they do, they tend to be informational ("what is a tort") rather than commercial ("personal injury lawyer near me").
• Financial services: 24% AI Overview rate. Anything involving specific financial advice or product recommendations is still served traditionally.
• Local home services: 15% AI Overview rate. Queries like "roofer near me" or "emergency plumber [city]" almost never trigger AI Overviews. The local pack still dominates.

How to get cited in AI Overviews

Being cited within the AI Overview is the new "position zero." When Google's AI pulls from your page to construct its answer, your site appears as a source link within the overview itself — and our data shows this actually increases CTR compared to the same organic position without any AI Overview present.

We tracked citation rates across all six client sites and reverse-engineered the factors that correlate with being cited. Here's what we found:

1. Structured data is non-negotiable

Pages with comprehensive structured data markup were cited 3.2x more often than pages without it. Specifically, FAQ schema, HowTo schema, and Article schema with author markup had the strongest correlation. This makes intuitive sense — structured data gives Google's AI a machine-readable way to extract and attribute information.

We saw the clearest example with our healthcare client. After implementing comprehensive FAQ schema on their condition pages (50+ questions per condition), their AI Overview citation rate went from 4% to 31% within six weeks. Traffic to those pages actually increased 14% despite the AI Overview being present.

2. Original research and first-party data

Google's AI strongly prefers citing pages that contain data not available elsewhere. Our B2B SaaS client published a survey of 2,400 project managers about remote work productivity. That single page gets cited in AI Overviews for 23 different keywords. Generic blog posts covering the same topics with no original data get cited for zero.

The pattern is clear: if your content can be fully synthesized from other sources, the AI Overview will synthesize it and not cite you. If your content contains unique data points, proprietary research, or original analysis, the AI has to cite you because the information doesn't exist anywhere else.

3. Entity authority and E-E-A-T signals

Pages written by authors with established Google Knowledge Panel entities are cited 2.7x more often than pages with no author attribution. This aligns with the E-E-A-T framework — Google's AI trusts sources connected to verified human experts.

Last month, a client asked us why their competitor kept appearing in AI Overviews while they didn't. Both sites had similar content quality and domain authority. The difference? The competitor's articles had named authors with LinkedIn profiles, conference speaking pages, and published research papers. Our client's content was attributed to "Staff Writer." We fixed the attribution, added author schema, built out author pages — and within 8 weeks, AI Overview citations increased 340%.

4. Concise, direct answer formatting

AI Overviews pull from content that provides clear, direct answers in the first 2-3 sentences of a section. We tested this with our e-commerce client by reformatting their buying guides. The original format buried the recommendation after 300 words of context. The new format led with a direct answer, then provided supporting context. Citation rate doubled.

The formatting pattern that works: direct answer (1-2 sentences) followed by supporting evidence (2-3 sentences with specific numbers) followed by nuance or caveat (1-2 sentences). This mirrors the structure AI Overviews use internally.

5. Content freshness and update frequency

Pages updated within the last 90 days are cited 1.8x more often than pages last updated over 6 months ago. Google's AI appears to have a strong recency bias — it prefers pulling from sources it considers current. We test this by monitoring citation rates before and after content updates. A simple update to add current data points and refresh the "last updated" date consistently improves citation rates within 2-4 weeks.

The AI Overview optimization strategy we use with clients

Based on 8 months of data, here is the exact strategy we now implement for every PxlPeak client. We call it the "Citation-First" approach because the goal is not to fight AI Overviews but to get cited in them.

Step 1: AI Overview keyword audit

We run every tracked keyword through SEMrush's SERP Feature filter to categorize them into three buckets:

• No AI Overview (yet): Traditional SEO optimization. Focus on rankings and featured snippets.
• AI Overview present, we're cited: Protect the citation. Keep content fresh, expand structured data, add more unique data points.
• AI Overview present, we're not cited: This is where the real work happens. Reverse-engineer which sources are being cited and what they have that we don't.

Step 2: Content restructuring for citation

For every page targeting a keyword with an AI Overview, we restructure the content following the direct-answer pattern described above. We also add:

• FAQ sections with schema markup (minimum 5 questions per page)
• Data tables with specific numbers, percentages, and named sources
• Named author with full bio, credentials, and author schema
• Clear H2/H3 hierarchy matching common question formats
• "Last updated" dates with a commitment to quarterly refreshes

Step 3: Original research investment

We recommend every client produce at least one original research piece per quarter. It doesn't need to be a 50-page whitepaper. A survey of 200 customers, an analysis of your own platform data, or a case study with specific numbers — any of these create citable, unique data that AI Overviews can't synthesize from elsewhere.

Our roofing client published a "2026 Long Island Roof Replacement Cost Report" based on their actual job data from the previous 12 months. That single page now gets cited in AI Overviews for 11 different cost-related queries and drives 340 organic visits per month — more than it did before AI Overviews existed on those queries.

Step 4: Author and entity building

Every piece of content gets attributed to a named expert. We build out author pages with:

• Full professional bio with verifiable credentials
• Links to published work, speaking engagements, and professional profiles
• Person schema with sameAs links to LinkedIn, Twitter, and Google Scholar
• Regular bylined content on external authoritative sites (guest posts, industry publications)

Step 5: Monitor and adapt

AI Overviews are not static. Google adjusts which queries trigger them, how they're formatted, and which sources they cite on a rolling basis. We check citation status weekly and run full audits monthly. When a citation is lost, we investigate immediately — it usually means a competitor published fresher data or Google changed the query interpretation.

The contrarian take: AI Overviews are good for real businesses

I know this is an unpopular opinion in the SEO community, but I believe AI Overviews are a net positive for legitimate businesses. Here's why.

Before AI Overviews, the SEO game rewarded content volume. Publish 500 blog posts targeting every variation of every keyword, get a ton of traffic, and hope that 1-2% converts. Most of that traffic was worthless — people looking for a quick answer who bounced immediately.

AI Overviews eliminate that game. The quick-answer seekers get their answer without clicking. The content farms that existed solely to capture those clicks lose their business model. What's left is higher-intent traffic that actually wants to engage with your business.

We tested this theory directly with our personal injury law firm client. They had 340 blog posts, most targeting long-tail informational queries like "what to do after a car accident in [state]." AI Overviews now answer many of those queries directly. Organic traffic to the blog dropped 31%. But — and this is the key — lead form submissions from organic search increased 8% in the same period. The traffic they lost was never going to hire a lawyer. The traffic they kept is exactly the traffic they want.

What to do right now: a 30-day action plan

Whether you're panicking about AI Overviews or haven't thought about them at all, here's the prioritized action plan we give every new client:

Week 1: Audit and assess

• Export your top 200 organic keywords from Google Search Console
• Check each keyword in SEMrush or Ahrefs for AI Overview presence
• Categorize into the three buckets (no AIO, cited, not cited)
• Calculate your "AI Overview exposure rate" — what % of your traffic-driving keywords have AIOs?

Week 2: Quick wins

• Add FAQ schema to your top 20 pages (use our free SEO checker to validate implementation)
• Update author attribution on all content — remove "Staff Writer" and add real names
• Refresh your top 10 pages with current data and updated "last modified" dates
• Restructure introductions to lead with direct answers instead of preamble

Week 3: Content gaps

• Identify 5 keywords where you're ranked but not cited in the AI Overview
• Analyze the cited sources — what do they have that you don't?
• Create or update content to match or exceed the cited sources
• Add unique data points, original statistics, or proprietary insights

Week 4: Foundation building

• Build out author pages with full bios, credentials, and schema markup
• Plan your first original research piece (survey, data study, or industry report)
• Set up weekly AI Overview monitoring in your rank tracking tool
• Brief your content team on the Citation-First content structure

Looking ahead: what's next for AI Overviews

Based on the trajectory we're tracking, here are our predictions for the next 12 months:

• Expansion to commercial queries: AI Overview presence on commercial queries will rise from 23% to 40%+ by end of 2026. Google is already testing product comparison AI Overviews with shopping data integration.
• Interactive AI Overviews: Google is testing follow-up question functionality within AI Overviews. This will further reduce clicks for multi-step informational queries but create new citation opportunities for each follow-up response.
• Ads within AI Overviews: Google has started testing sponsored links within AI Overviews in limited markets. When this rolls out broadly, it will create a new paid channel but also a new competitive threat to organic visibility.
• Source quality scoring: Google will increasingly differentiate between sources based on E-E-A-T signals. Sites with verified expertise will see more citations; generic content sites will see fewer.

The sites that start adapting now — investing in original research, building author authority, implementing comprehensive structured data — will be in a strong position regardless of how AI Overviews evolve. The sites that keep publishing generic content and hoping for the best will find themselves increasingly invisible in search results.

AI Overviews are not going away. They're going to expand. The question is not whether they will affect your business — it's whether you'll be cited in them or buried underneath them. We know which side we want our clients on.