How AI Overviews are changing who gets found in search

More web searches now get answered before anyone clicks a single link. Google builds an answer, called an AI Overview, straight into the results page, pulling together information from a handful of sources it trusts enough to quote directly.

For some businesses, that presents a genuine opportunity. For others, it’s one more way to become invisible.

We’ve written before about the mistake of treating robots and humans as rivals for your attention, when the real job is writing content that’s easy enough for the algorithms to crawl and understand, and good enough for the person to value it once they do. After all, good SEO is good GEO (generative engine optimisation) and vice versa, because the algorithm and the reader are both looking for content that’s worth their time.

This article builds on the thinking we put forward in our previous blog, ‘Good SEO was never just about the robots… or the humans‘. You should check it out, if you haven’t read it already.

In this piece, we’re going to go a bit deeper. We’ll look at what happens when Google builds an AI Overview, what gets a business cited inside one, where a lot of businesses are currently wasting their effort, and what’s at stake for the ones that get this wrong. If you’re already investing in the fundamentals, this is where that investment starts to pay off. If you’re not, this is why the gap is likely to matter more than it used to.

A short history

You’ll have seen SEO advice shift more than once over the past decade or so, and it hasn’t moved in a straight line from one big idea to the next. Keyword density was the whole game in the early days, until a wave of algorithm updates started penalising thin, spammy content rather than rewarding whoever repeated a keyword the most.

From there, Google’s algorithms got steadily better at understanding the meaning behind a search rather than just matching the words in it, learning to handle context, intent and natural language rather than strings of keywords. Technical changes also started to carry more weight, from mobile-first indexing to Core Web Vitals rewarding fast, stable, usable pages.

Somewhere in among all of that, Google introduced E-E-A-T, the guidelines its human reviewers use to rank content.

E-E-A-T isn’t a ranking factor in itself. It’s the framework that the people who assess search quality use to judge whether content deserves to rank. However, it shapes how Google’s algorithms get trained and refined.

AI Overviews are the latest layer on top of all of that. They’re built on the same index, the same quality signals and the same underlying algorithms as everything that came before. You don’t need a different kind of content, written for a different kind of reader, to work with them. However, you do need content that clears a bar which keeps rising, because AI Overviews are more selective about what they’ll quote than the traditional SERPS results ever were.

So, let’s look at how AI Overviews decide what to include.

What happens when Google builds an AI Overview?

An AI Overview isn’t a separate search engine running its own rules. It pulls its answers from the same index and the same core ranking systems that power traditional search results. That’s why a page that already performs well in search tends to be a stronger candidate for citation, too.

The difference lies in how Google assembles the answer. Its algorithm uses a technique called fan-out to expand a single search into several related queries behind the scenes, pulling together a broader and more varied set of sources than the traditional search results would ever show on one page of results. Ask a question about, say, choosing a commercial fridge for a small catering business, and Google might check its indexed sources on energy efficiency, servicing costs, capacity planning and installation requirements all at once, then weave the most useful parts of each into a single, AI-generated answer.

That matters because it changes how many chances your content gets to be useful. On a traditional results page, one strong page on one topic can earn its place. Inside an AI Overview built from a fan-out of related queries, your content might be competing to answer a narrower slice of a bigger question, alongside sources that specialise in exactly that slice. So, a page that covers a topic broadly but without enough depth is less likely to be pulled in than a thinner page that answers one part of it properly.

That has a direct consequence for how you write. Being generally good isn’t enough if your content only answers part of the question. The pages most likely to get pulled into an AI Overview tend to share a few traits:

• Topic sentences that state the key point clearly upfront
• Paragraphs that make sense when read in isolation rather than only in the context of the whole page
• Factual claims backed by evidence

None of that means chopping your content into keyword-stuffed fragments. It means writing clearly enough that a single paragraph can stand on its own and still be useful, whether it’s read as part of your page or lifted into an AI-generated answer.

Where businesses waste effort chasing the wrong signals

Google has been fairly direct about what doesn’t move the needle in search. Yet these are exactly the things that many businesses keep reaching for, because the pace of change makes people feel like they need to be doing something new, even when the old advice still holds.

LLMs.txt files, for example, don’t affect how Google Search treats your content, whatever their name might suggest.

Keyword-stuffed copy dressed up as ‘AI-friendly’ content performs no better than the keyword-stuffed copy of five years ago, because Google’s language models now read for meaning, not repetition.

Artificially chunking your content into short, disconnected snippets in the hope of making it more ‘extractable’ tends to make it harder to read and no more likely to be cited.

And chasing inauthentic mentions across forums and low-quality sites to create the appearance of authority carries the same risk it always has. Google’s systems are built to recognise genuine trust signals, not manufactured ones.

None of this is really about AI Overviews specifically. It’s the same pattern that’s played out with every shift in search over the past decade. A new shift appears, businesses assume the rules have changed, and effort gets diverted towards tactics that were never going to work in the first place.

So, what does move the needle?

The list of things that work will sound familiar, because it’s the same list we included in our previous blog (see here), namely:

• Author attribution, so a real person, not an anonymous ‘admin’ account, stands behind what’s published.
• First-hand expertise, so you can write about the things you’ve actually done, rather than only researched.
• Solid technical fundamentals, a fast, secure, mobile-friendly website that doesn’t get in its own way.
• Content structured clearly enough that both a person skimming on their phone and an algorithm parsing the page can follow it.
• Genuine trust signals, including accurate information, transparent source attribution and corrections made openly when something needs fixing.

We covered the detail of E-E-A-T properly last time, so we won’t repeat it at length here. But the point worth restating is simple. Nothing about AI Overviews changes what earns trust. It just changes how that trust gets rewarded.

Why the gap is widening, not narrowing

Just a few years ago, a traditional results page might have shown ten ‘blue links’, giving a reasonable spread of businesses a chance to be seen, even from a middling position. Someone might scroll past the top result and still land on yours.

AI Overviews draw from a much smaller number of sources to build a single, credible answer, and there’s no scrolling past it to find you further down. Being one of ten isn’t the same proposition as being one of three or four.

That means the businesses already investing in building genuine authority with their content are becoming more visible as a direct result of that investment, not just holding steady. Every piece of properly attributed, well-researched content adds to the case Google’s systems are building for treating that business as a trusted source. And that case only gets stronger with each new piece, rather than resetting to zero.

On the flip side, businesses that don’t invest in content that builds authority aren’t just staying where they’ve always been. They’re becoming harder to find, because the visibility that decent, mid-table content used to earn is increasingly reserved for the sources Google’s systems are confident enough to quote directly.

The knock-on effect is that as more searches get answered inside AI Overviews without a click ever happening, the value of being one of the trusted sources behind that answer increases, even if it doesn’t always show up as a visit to your website. Being cited by name inside an AI Overview builds the same kind of recognition as a mention on a trusted website, regardless of whether the reader clicks through immediately. This gap is widening faster than it did during previous shifts in search, because AI Overviews consolidate attention onto fewer sources by design.

If you want to see where you currently stand, Google Search Console has begun rolling out reporting that shows the number of impressions your content is receiving specifically from AI features on Google Search. If you’re a retailer, Merchant Centre has equivalent reporting showing how your product listings appear across Google’s generative AI features. Both are worth checking now, while there’s still time to act on what you find, rather than waiting until a competitor has already claimed the ground.

What should you do right now?

Start by checking your Search Console data, if you have access to it, to see whether your content is showing up in AI features at all, and which pages are earning that visibility.

Look at whether your published content is attributed to a person, or whether it’s sitting under a generic byline that tells nobody, human or algorithm, anything about who’s behind it.

Test a handful of your most important pages by asking an AI tool to summarise them in a couple of sentences. If it struggles to produce a clear, accurate summary, or generates details you never stated, an AI Overview built from your content is likely to struggle in the same way.

Look, too, at whether your best content answers a question fully, or only touches on it. A page that mentions a topic in passing is a weaker candidate for citation than one that answers it properly, even if the passing mention sits on an otherwise strong page.

Treat whatever gaps you find as priorities, because the businesses closing these gaps now are the ones AI Overviews will increasingly be built around.

It’s also worth checking how your competitors are showing up. If you search for the questions your customers are likely asking and see the same two or three businesses cited repeatedly inside AI Overviews, that’s not a coincidence. It’s usually a sign that those businesses have been doing this work consistently for a while, which gives you a realistic benchmark for what ‘good enough to be cited’ looks like in your sector.

How can QBD help?

None of this is a reason to panic. However, it does call for the kind of steady, deliberate SEO work that’s always separated the businesses that perform well in search from those that don’t. It means doing the fundamentals properly and consistently enough that the algorithms evaluating them take notice.

If you’d like a clear picture of where your content currently stands, we can help. QBD Pulse, our structured, strategic, ongoing digital marketing programme, can take care of this every month for you.

We’ll create SEO blogs, social posts and email campaigns that highlight your expertise, build genuine authority and perform for your readers and for the algorithms deciding what gets seen. Your monthly Pulse Score then tracks how your subscribers are engaging with your content, so you know what’s landing and who’s ready for a conversation.

Get in touch with our team today, and we’ll show you what’s holding your content back, and what closing that gap could look like for your business.