How to Rank in AI Overviews and What Google I/O Got Wrong About the Risk

Most of my clients asked me questions this month.

Not about whether SEO is dead. About AI's impact on their brand — on whether their content still reaches the right people, whether they're still visible where it matters. Some of it comes from fear. But there's something healthy in it too. They're watching. They're trying to understand. They want to act.

What none of them needed was a consultant posting "search is over" the morning after a Google keynote. What they actually needed was a practical answer to how to rank in AI Overviews — and why the panic narrative makes that harder to find.

Is SEO dead after Google AI mode?

Let me be direct: posts like that are clickbait. We live in a world where everyone wants their fifteen minutes — where a hot take gets more reach than a measured analysis — and "SEO is dead" is the most reliable hot take in this industry. It comes out after every major Google announcement, every algorithm update, every interface change.

Search has evolved since its creation. It will keep evolving. It's not going anywhere.

What Google I/O actually announced was not the end of web results. Google confirmed the opposite — traditional results remain, blue links still exist, and the Web tab still works. What changed is that AI answers and agents now occupy more of the default experience. The blue links are still there. They're just less front and centre than they used to be.

That's a real shift. The risk is real. But the people naming it as "the death of search" are wrong — and so is Google's line that nothing has changed. The actual risk sits between both positions, and it's economic, not existential.

What the traffic data actually shows — and what most coverage missed

Here's what I've observed across my clients this year, and it's more interesting than the headline numbers.

Yes, we've seen impression drops. Yes, click drops. But funny enough — not conversion drops.

The people willing to buy are still finding these brands, because the content replies to a real need. Where we've seen the sharpest declines is in high-level, top-funnel pages — the kind of page that exists to answer a simple informational query. And honestly? You don't need a 1,200-word post to answer a simple question. AI can do that. It should do that. That's not a loss worth grieving.

The content that's holding, and in some cases actually gaining clicks and conversion, is the content built around genuine pain points. The specific problem a customer has been carrying for six months. The question they've been too embarrassed to ask their accountant, their builder, their lawyer. That content is not being absorbed by AI summaries. It's being clicked on by people who have a real decision to make.

A field study published via Search Engine Journal found that AI Overviews reduced organic clicks by 38% on triggered queries — but no change in user satisfaction. That pattern is real. What the coverage missed is that it lands unevenly. Informational content takes the hit. Intent-driven content, the kind built around what someone actually needs to resolve, holds.

The information agent question nobody is asking cleanly

The detail from Google I/O that deserved more serious attention than it got: information agents.

These are Google features that monitor the web on a user's behalf and surface synthesised updates inside Google — without the user ever visiting your site. Your content is consumed. No visit. No session. Nothing in your analytics.

I'm not the biggest fan of where this is heading, and I want to say that plainly. The short answer function is useful. There's a real convenience argument. But I find myself asking a question I haven't seen others ask: how are they going to filter that? Who decides which sources get synthesised and which get left out?

There's a freedom of speech dimension to AI-mediated content delivery that the industry is not talking about seriously enough. When an agent decides what synthesis to surface to a user, it is making editorial decisions. At scale. Invisibly. That's something we need to be aware of — not as a reason to panic, but as a reason to keep building the kind of content authority that makes you impossible to omit.

When those filtering decisions favour established, high-volume sources over specific, deeply human work, it's not just an editorial concern — it's a revenue one.

Glenn Gabe, SEO consultant at G-Squared Interactive, named the commercial risk clearly: "For publishers, information agents can hit ad revenue big-time as less people will be visiting websites." The sites most exposed are those whose content answers ongoing monitoring needs — availability, pricing, regularly updated information. If that's your business, this is the shift to watch.

For everyone else, the practical response is cleaner schema markup and structured layouts — FAQs, How-to content, LocalBusiness data — so that when agents do pull from your content, they attribute it correctly rather than absorbing it anonymously.

How to rank in Google AI Overviews

Let me give you a concrete example from my own brand.

I rank for "AIO SEO consultant Sydney." It's precise. It's geographic. It's exactly what someone types when they've made a decision and need to find the right person. I do not rank for "GEO consultant Australia" — that term is too broad, too generic, too disconnected from what I actually do and who I actually serve.

That distinction is the whole answer.

The clients of mine who are showing up in AI search are the ones who have the suburb in there, the specificity, the example of a real outcome and why it happened. Not because they've optimised their way there with a clever technical fix, but because their content is actually about something specific. It could only come from them. There's no interchangeable version of it sitting on a competitor's site.

Content built this way gets cited rather than absorbed. AI Mode needs to show its source when it quotes expertise it can't synthesise from a pool of identical pages. Entity authority — the kind built on specificity, geography, demonstrated outcomes — is what earns those citations.

The interface around search has changed. The underlying logic has not.

Before you change anything: a short audit

• Which pages are built around specific customer pain points — a problem someone has been living with, a decision they need help making? Those are your safest pages. Invest more in them.

• Is your content geographically or categorically specific? "AIO SEO consultant Sydney" is a different category than "digital marketing consultant." Specificity is not a limitation — it's a targeting strategy.

• Do your pages carry real examples — a client outcome, a specific scenario, a case that demonstrates your point? Generic claims won't earn citations. Specific ones do.

• Review your last 10 content briefs. Were they built keyword-first or need-first? The first approach produces content AI can absorb. The second produces content AI has to cite.

• Identify your 10 highest-traffic pages. Which ones answer simple informational queries? Those compete with AI summaries now. Don't panic — but understand they are your most exposed pages.

• Check whether you have schema markup on article and FAQ content. This helps AI features identify your content as a citable source rather than background noise.

Google AI mode and SEO strategy — what to actually do now

Here's the thing about the "be number one" mentality that Google I/O exposed as fragile: you don't need to be number one in everything.

You need to be number one where your clients are actually coming from.

If your ideal customer is a small business owner in a specific city, searching for a specific type of help, at the moment they've decided to act — that's the search you need to own. Not every search in your category. Not the broadest possible term. The one that sends you the person who is ready to buy.

AI search has made this more true, not less. Broad informational content is getting absorbed. Specific, geographically rooted, outcome-oriented content is getting cited. The businesses that will feel this shift hardest are the ones who built their visibility on volume keywords rather than specificity.

It's a puzzle — technical, content, and authority all working together. But the piece most businesses are underweighting is not technical. It's the willingness to be specific. To put the suburb in. To say what they actually do for a real kind of person with a real kind of problem, rather than staying broad enough to theoretically appeal to everyone.

Put salt in your content.

The anecdote from a real client conversation. The suburb. The outcome and how it happened. The opinion you hold that your competitors are too cautious to publish. The kind of specificity that makes AI say: this has to be cited, not summarised.

No prompt generates that. It comes from knowing what you actually know — and being willing to say it.

Google I/O accelerated a shift that's been underway for two years: AI answers are replacing clicks on content that doesn't need a human to have written it. That's the real risk. And the answer is the same as it's always been.

Your content either tastes like something. Or it doesn't.