AI Overviews are killing your click-through rate. Here is what to do about it.

The headline number that is freaking everyone out

Google AI Overviews now appear in 25.11% of all searches (Conductor 2026 benchmark). In healthcare, that figure is 48.75%. Ahrefs has measured a 58% drop in click-through rate on top-ranking pages when an AI Overview is present, up from 34.5% the year before.

Most agencies are reading this as a doomsday signal: organic search is dying, our clicks are gone, the channel is broken. That is the wrong read. The channel is not dying. It is splitting into two distinct outcomes:

1. Pages cited inside the AI Overview, which capture brand impression and a smaller pool of high-intent clicks.
2. Pages ranking organically but not cited, which still get clicks, just fewer of them.

You can be on either side of that split. The strategy you choose decides which.

What actually happens when an AI Overview appears

When Google decides a query deserves an AI Overview, it does two things in parallel. It runs the traditional ranking algorithm and surfaces the usual ten blue links underneath. Then it sends those top-ranking pages plus a wider retrieval set to its generative model, which writes a 2-4 paragraph answer and chooses 3-7 sources to cite by name and link.

The user reads the Overview. About 40% click through to one of the cited sources. The rest get their answer and move on, or refine the query.

That means three commercial outcomes for any page you own:

• Cited inside the Overview: your brand gets seen by everyone who saw the Overview (impression), and you capture roughly 40% of the click flow split among 3-7 sources.
• Ranked top 10 but not cited: you keep your traditional CTR but it is now diluted by Overview-driven query abandonment.
• Not ranked: you are invisible to both the Overview and the click stream.

Being cited is now a separate game from ranking. The pages that win citations have measurably different structure to the pages that just rank.

The five things AI Overview cited pages share

Looking at pages that get cited consistently versus pages that rank but do not, five patterns repeat:

One: a direct answer to the query in the first 40 to 60 words of the page, written as a standalone paragraph that makes sense lifted out of context.

Two: question-led H2 headings that mirror how people actually phrase prompts ("How much does Google Ads management cost in Brisbane?" not "Brisbane Google Ads pricing").

Three: specific statistics with named sources, inline. Generic claims do not get extracted. Cited claims do.

Four: structured data, especially FAQ schema and Article schema with verified author bios. These help Google parse what is an answer and who is making it.

Five: freshness. Pages updated within the last 60 days are 28% more likely to be cited than older pages on the same topic (Conductor 2026).

None of these are radical SEO. They are all extensions of patterns that have worked since featured snippets. What is new is how heavily AI Overviews weight them.

What to do if your pages are being out-cited by competitors

Five-step playbook that has moved citation frequency on real Brisbane client sites:

1. Pull your top 20 highest-value pages (the ones that drive leads, not just traffic) and identify the primary query each one targets.
2. For each primary query, query Google directly and screenshot the AI Overview if one appears. Note who is cited. If your competitors are cited and you are not, that is a fixable gap.
3. Rewrite the opening 40-60 words of each page to directly answer the query in a standalone paragraph. No build-up, no context-setting, just the answer.
4. Add an FAQ block on each page covering 4-6 of the most common related questions, with FAQ schema marked up.
5. Add specific statistics with named sources wherever you make a claim. Numbers get extracted, vague claims do not.

This is six to eight hours of work per page. The lift in citation frequency typically shows up within 14 days, because Google re-crawls and re-evaluates citation candidates on a roughly weekly cycle for active pages.

What this changes for measuring SEO success

The metrics that mattered five years ago (rank, sessions, click-through rate) are still useful, but they are no longer sufficient. The three metrics worth adding to your monthly report:

• AI citation frequency: how many of your target queries cite your brand inside ChatGPT, Perplexity, Google AI Overviews and Copilot. Sample monthly with the same query list.
• Share of model: against your top 3-5 competitors, what percentage of citation opportunities go to you versus them?
• AI referral traffic in GA4: Conductor and Ahrefs both report AI-sourced visitors convert at meaningfully higher rates than traditional organic. Track them separately.

These metrics do not replace traffic and conversion data. They sit alongside it, and they catch shifts that traditional reporting misses entirely. A site can be losing traditional traffic while gaining citation visibility, which means losing low-intent clicks while keeping or growing high-intent leads. That is a win, but only if you are measuring the right things.

The honest summary

AI Overviews are not killing organic search. They are reshaping it. The strategy is not to fight the format, it is to become the source the format cites. The work to do that is concrete, measurable, and within reach of any business with 20 high-value pages and 6 to 12 weeks of consistent execution.

If you want to baseline where your brand currently sits across the major AI engines, our GEO Audit tests 20 of your target queries and gives you a prioritised list of fixes. Or book a strategy call and we will walk through which of your existing pages are closest to citation-ready.