The 34% Problem: What Happens to Your Website When AI Answers First

Pull up your Google Analytics for the past twelve months. Look at organic traffic. If you are a service business, there is a good chance the line is flat or trending down.

The usual explanation is that competition got tougher, or that SEO is harder than it used to be. Both of those things are true. Neither of them is the main reason.

The main reason is that Google started answering your customers' questions before they could click on your website.

34% was just the beginning

In April 2025, Ahrefs analyzed 300,000 keywords and found that Google's AI Overviews reduced clicks to the top-ranking result by 34.5%. That was the first large-scale measurement. The industry noticed. Then the number kept getting worse.

By December 2025, Ahrefs ran the same study again. The decline had worsened to 58%.

Seer Interactive tracked 3,119 informational queries across 42 organizations over fifteen months. Organic click-through rates on queries where an AI Overview appeared dropped from 1.76% to 0.61%. That is a 61% decline. Paid click-through rates fell 68%.

These are not edge cases or cherry-picked datasets. This is a trend line with one direction.

What actually happened

In May 2024, Google began rolling out AI Overviews: a generated summary that appears above the organic results, synthesized from multiple sources, designed to answer the query without requiring a click. By early 2026, AI Overviews appear on roughly half of all U.S. Google searches, with the number continuing to climb.

The mechanic is simple. A homeowner searches "how much does a new AC unit cost installed." Three years ago, she would have seen ten blue links and clicked two or three of them. Today she sees a generated paragraph at the top of the page that synthesizes pricing data from multiple sources, with citations in small text. She reads it. She has her answer. She does not scroll down. She does not click.

Your website might be one of the sources cited in that overview. It might even be the primary source. But the click never happens. The user got what they needed without leaving Google.

Similarweb's data tells the broader story: zero-click searches have climbed from 56% to 69% since AI Overviews launched. On mobile, it is 77%. When Google's dedicated AI Mode is active, 93% of searches end without a click.

This is not the same as zero-click search

Google has been eating clicks for years. Featured snippets, knowledge panels, People Also Ask — all of these reduced click-through rates incrementally. SEOs adapted. They optimized for Position Zero. They restructured content to win the snippet.

AI Overviews are different in kind, not just degree.

A featured snippet pulls a single block of text from one page. If the user wants more, they click through. An AI Overview synthesizes information from multiple sources into a coherent answer. It doesn't excerpt your content. It replaces the need to visit it.

And it is getting better at this. Fast. The coverage has expanded from a narrow set of informational queries to nearly every vertical. Healthcare queries trigger AI Overviews 88% of the time. Education, 83%. B2B Technology, 82%. The expansion into service categories is accelerating.

For local service businesses, there is a temporary reprieve: pure local-intent queries like "plumber near me" still mostly return the traditional local pack. But the moment a query adds informational intent — "how much does a tankless water heater cost" or "is a heat pump worth it in Texas" — the AI Overview takes over. And those informational queries are the top of your funnel. They are how homeowners start the research process that ends with hiring someone.

The measurement problem

Here is what makes this particularly dangerous: your analytics cannot tell you it is happening.

Google Analytics shows impressions, clicks, and CTR for organic search. If your impressions are stable but clicks are declining, you are seeing the AI Overview effect. But most business owners and their agencies look at traffic, not CTR. Traffic is down 15%? Must be time to invest more in SEO. Or run more ads.

The diagnosis is wrong. The traffic did not go to a competitor. It did not go anywhere. It was absorbed by a generated answer on the search results page. Spending more on SEO to rank higher does not fix this, because even Position 1 is losing 58% of its clicks when an AI Overview is present.

This is happening even if your rankings have not changed. You can hold Position 1 and still lose more than half your clicks. The problem is not your rank. The problem is that Google put an AI-generated answer above it.

Two problems, one cause

If you have been reading this blog, you know we have written about a parallel channel — the growing share of consumers who skip Google entirely and ask ChatGPT or Perplexity to recommend a service provider. That is the search that never reaches Google.

This post is about the search that does reach Google but never reaches you.

They are two symptoms of the same underlying shift. AI systems — whether Google's own or third-party — are answering questions in ways that absorb the click. The result for your business is identical: fewer people landing on your website, with no record in your analytics of why.

And they compound. Your top-of-funnel informational traffic declines because AI Overviews answer the question on Google. Your bottom-of-funnel "hire someone" traffic declines because a growing share of consumers ask ChatGPT instead. The squeeze comes from both ends.

Who wins in this environment

There is a counterintuitive finding buried in Seer Interactive's data. Brands that are cited in AI Overviews earn 35% more organic clicks and 91% more paid clicks than brands that are not cited.

Read that again. If your business is a source that the AI Overview draws from, you do not lose traffic. You gain it. The AI Overview acts as an endorsement. Users see your brand in the citation and click through with higher intent than a standard organic result.

This flips the problem on its head. The question is not "how do I fight AI Overviews." It is "how do I become the source they cite."

The same pattern holds across AI systems outside of Google. Ahrefs' study of 1.4 million ChatGPT prompts found that 76.4% of the most-cited pages had been updated within the last 30 days. Recently updated content receives 3.2x more citations than older material. Content freshness is not a nice-to-have. It is the single strongest predictor of whether AI systems — Google's or anyone else's — use you as a source.

The businesses that show up in AI results share two traits:

1. Their data is structured. Not marketing copy. Discrete, parseable facts: what they do, where they operate, how they perform, what credentials they hold. Formatted in Schema.org JSON-LD, not buried in paragraph text.

2. Their data is fresh. Updated within weeks, not years. AI systems have a measurable preference for recently verified information. A page last touched in 2023 is functionally invisible to most AI citation engines.

What to do this week

Check your Google Search Console. Filter by queries where you rank in the top 5. Compare CTR year-over-year. If CTR has declined significantly on queries where your ranking has not changed, AI Overviews are absorbing your clicks.

Then check whether your website has Schema.org LocalBusiness markup in JSON-LD. Not just name, address, phone — actual operational data. Services offered. Service area. Credentials. Years in operation. Most local business websites have none of this, or have broken markup auto-generated by a plugin that no one has audited.

Then ask the harder question: when was the last time the structured data on your website was updated? If the answer is "when we built the site," you have a freshness problem. AI systems that prioritize recently updated content are not finding yours.

This is the problem we built TrueSignal to solve. A TrustRecord is a structured, verified, machine-readable record of your business — operational metrics, credentials, service area, service mix — pulled from your actual systems of record and refreshed monthly. Three layers: server-rendered HTML, Schema.org JSON-LD, and canonical JSON. It is designed to be the source AI systems cite.

No one can guarantee that any AI system will cite any particular business. But the data is clear about what gets selected: structured, fresh, verified information. And it is equally clear about what happens to everything else.

Somewhere right now, a homeowner is Googling a question about a service you provide. Google is answering it with an AI Overview that synthesizes three sources. You are either one of those sources, or you are below the fold watching your CTR decline another point. The 34% problem only gets worse from here.