Search and SEO are changing, but don’t throw the baby out with the bathwater

Yes, AI Overviews are multiplying, clickthroughs are diminished, and brands are having to revisit their plans to strategise for mentions and citations. But good content still looks the same.

Search is changing and there is a heck of a lot of chatter about it. These two things seem pretty evident.

It doesn’t matter if you’re an AI power-user or a sceptical bystander who prefers artisanal copy – everybody’s search habits have no doubt changed in the last year or two.

The chatter from the SEO and marketing industry is understandable, because this thing is real. The numbers are changing thanks to the impact of Google’s AI overviews and, less so, AI assistants such as ChatGPT.

But let’s not panic. After all, debate in the world of search marketing is nothing new. For over two decades, there’s been a healthy back and forth between Google and the industry. And though I wouldn’t call SEO an arms race, brands have had to react to many algorithm updates over the years to stay ahead of their competitors (depending on your age, reader, you might get a waft of nostalgia from the words Florida, Penguin, Panda or Hummingbird).

So, before marketers get overwhelmed by scores of new playbooks for GEO or AEO (choose your favourite term), let’s ask some straightforward questions:

• • • • • How is search and search behaviour changing?

        • What impact is this having on brands?

        • What should they do about it, practically?

        • And what mindset shift is required for the future of search?

How is Google search and search behaviour changing?

AI overviews are proliferating

At the end of 2024 I wrote in Marketing Week about how I had noticed my own behaviour changing. I was relying on Google’s AI overviews (AIOs) for many low-jeopardy informational searches such as ‘how to reheat McDonald’s chicken nuggets’ (don’t judge). And I wondered what this could mean for brands (plus ça change, informational content should be authoritative, unique, and targeting in-market audiences; and anything else should have some brand-building personality).

Since last year, though, there’s a general consensus that AI summaries have increased dramatically. There are a whole host of studies trying to pin down how often they appear, with some citing the figure of 50% of all searches (as quoted by McKinsey).

And though a new report this month from Ahrefs gives a more conservative figure of 21% of all keywords, that’s still up from 9.5% in May, and there are some interesting nuances in the November data, too:

• • • • • AIOs are nearly twice as common for non-branded keywords than branded ones.

        • AIOs appear for 57.9% of all question queries (note here that recent Semrush research has shown the fondness of US searchers for simple ‘Is it good?’ or ‘Is it bad?’ style queries – it’s worth reading the summary just for the quotidian poetry in some of the top searches).

        • The categories with highest AIO share are Science (43.6%), Health (43.0%), Pets & Animals (36.8%), and People & Society (35.3%).

        • 44.1% of medical queries are in the so-called ‘your money or your life’ (YMYL) category, i.e. topics that could impact your health or your finances, trigger an AIO.

Anecdotally, these numbers feel about right to me, and explains how in the past twelve months my own AIO usage has inexorably spread, notably into ‘Dr Google’ territory. Where once I would scroll down to an NHS link, now I feel more or less comfortable, in the first instance, with an AI summary.

Clickthroughs have diminished where AIOs are present

This shouldn’t be a controversial claim. AIOs are designed to provide an answer there and then, not to send you off to a website that has the answer you need.

A recent Ahrefs study found that when present on a results page, AIOs reduce clicks by around 34.5%.

Google, however, argues that clickthroughs are more or less steady, but that crucially AIOs are sending a higher volume of quality clicks to websites, i.e. searchers that linger for longer and don’t immediately bounce. Indeed, there are companies such as Buffer and Ahrefs which have confirmed that visits from AI search, namely LLMs such as ChatGPT, are demonstrating much higher conversion rates for their own products.

The bottom line for a lot of website owners and publishers is that clickthroughs seem to have taken a hit.

SEOs have been sharing their Google Search Console performance reports, showing how clicks and impressions begin to diverge around April 2025, with clickthroughs dropping off as the proportionality constant (i.e. the clickthrough rate) diminishes. One SEO posting on X eloquently described this phenomenon as ‘The Great Decoupling’.

We’ve seen this dynamic before with Google’s featured snippets, the original ‘position zero’. And with AIOs and their citation links (where the supposedly quality clicks come from), there’s no way to isolate these particular clicks and measure them in Search Console. So, for now, this is a bit of a grey area.

Mentions, in particular, and citations bring a little mindset shift to SEO

Even casual searchers will have noticed the way that AIOs seem to work, drawing from a range of sources and summarising (generating) a best-of-all-worlds answer. This is done with a RAG system – the LLM is grounded using information sourced from across the web.

RAG means that your brand mentions in AIOs could come from your own content, but more likely, will come from third-party sources such as social media, reviews or publishers.

Mentions, therefore, bring all that lovely PR and brand building work into the glare of position zero. Where once, link building was a major driver of so-called SEO PR, in order to build website authority, now a backlink is no longer the be all and end all of media coverage. And indeed, mentions can be drawn from far broader sources than simply traditional media outlets.

Of course, ranking well in search correlates with LLM mentions. But mentions are quite obviously also a product of a famous or well-loved (or talked about) brand i.e. if you’re Marks & Spencer, you’re likely to rank well for certain clothing search terms, and get mentioned in an AIO about ‘quality British high street clothing’ (I just tried it out, and sure enough…). But smaller brands can also stick their elbows out here, gaining mentions from favourable Reddit discussions, for example.

Then, aside from mentions, there’s the citations in AIO results. The little source links and informational cards. These are driven by authoritative content (think Google’s E-E-A-T criteria), not dissimilar to traditional SEO. The RAG framework relies on crawling your site, hoovering up that helpful schema markup as usual.

Semrush has a good line for the two-pronged strategy brands need to approach mentions and citations: “The sentiment battle” (discoverability, being talked about, the brand’s community voice) and “the authority game” (authoritative sources that LLMs can easily digest).

Both approaches are needed because mentions do not equal authority. Only a minority of brands are able to consistently appear in mentions and citations.

AI search is an entry point for consumers

AIOs and other answer engines are simplifying complex searches. They are concertinaing the initial stages of many longer research journeys.

Think of a traveller planning their next getaway. The seeking out of potential destinations that may suit a diver / hiker / skier / raver might begin in Google’s AIOs or ChatGPT, with the searcher asking a series of questions to land on some options that suit their price range, itinerary and tastes. Once the search is narrowed down, the user in question might jump to the traditional search engine results as they look for specific hotels, tourist boards and airlines.

In this respect, searchers may arrive at a travel site ready to book. Instead of hitting a destination specific landing page from an informational non-branded search, they might have searched for a specific branded term and jumped into site search already knowing what they want.

Or, to look forward to the expanding future of agentic commerce, shoppers may even have used Instant Checkout in ChatGPT to buy single items without leaving their chat.

ChatGPT usage isn’t necessarily cannibalising search

Finally, on the changing nature of search, it’s worth pointing out that despite many bullish predictions of users shifting over to chat interfaces, at the moment there’s no data to suggest a doomsday scenario for Google.

Recent data has shown that new users of ChatGPT actually start to use Google Search more often as a result. AI tool adoption may be growing rapidly, but there’s no sign of people abandoning Google search.

What impact are AI overviews having on brands?

Well, some are losing traffic. Publishers in particular are wrestling with this issue, with lawsuits abounding in several different countries.

Just as social media takes eyeballs away from the wider web, Google’s AIOs may do the same. If you’re a user, you don’t really care, you want the answer to your question, not to land on an ad-stuffed website.

It’s a familiar story, one of steadily increasing competition for people’s attention. Where once, kids would read the breakfast cereal box or commuters would read The Metro front to back, for want of anything better to do, now they have almost infinite choice.

Sundar Pichai at Google has said that the number of web pages crawled has gone up by 45% in the last two years. Combined with the endless scrolling of social media, brands and publishers must arrest people’s attention before they get a chance to tell the full story.

The challenge of cutthrough explains why the creator economy is booming.

And now that earned content has a more prominent place in position zero, brands are realising they need more holistic strategies. If they can’t track clicks, they may need new measurement models, too, that better account for non-click engagement in search. 

So, what should brands do to go from SEO to GEO?

We’ve touched on some of these solutions or strategies already, but let’s put a list together.

• • • • • Conduct a gap analysis. Where are you being outcompeted in AIOs and answer engines? Across what topics?

        • Audit how your brand is being discussed in AI results. Does it match up to what you’re aiming for? If not, which sources may be responsible?

        • Research frequently cited publications and sites. Can you create a hit list for your brand to target?

        • Strategise for third party mentions. Are people talking about your brand on forums and in online communities? Can you build links with industry publications? 

        • Track these wider web mentions and those of your competitors. How do they change over time? Where are they being won and lost?

        • Use your subject matter experts as your biggest SEO asset. How can you incentivise and train people to share their knowledge publicly on platforms like Reddit, Quora, and industry forums?

        • Track traffic from AI assistants such as ChatGPT. How is it changing?

        • Continue to prioritise people-first content (E-E-A-T) for best chance of citation. What new and unique insights, data or experiences can you share?

        • Create easy to parse content. Can you include FAQs, bullet points, tables, conclusions, takeaways, subheadings, appropriate markup, semantic triggers (e.g. “Step 1”) or AIO friendly formats such as product comparisons or step by step guides? (Microsoft has some good guidance here).

        • Create accessible and user friendly content. No change there, right?

        • Prioritise recency in your content, all the better to get picked up by that RAG framework. Are you relying too heavily on evergreen assets?

        • Explore multiplatform publishing if you’re not already. Can you create or encourage content on YouTube and Reddit, two heavily influential and cited sources?

Looking at this list, you can see why I’ve mentioned not throwing the baby out with the bathwater. There are some new lines of enquiry for SEOs, new reporting if they haven’t started already, more PR style outreach and UGC wrangling, but on the whole, what’s good for standard Google search is also good for AIOs.

As CTI Digital’s Head of Content Sacha Crowther wrote after attending Brighton SEO earlier this year:

“Good content that is helpful, well-structured and authoritative should come out on top, regardless of whether it is cited in an AI response, or shared by humans. Content creators just need to think strategically and resist the urge to dramatically change how they work. Unless, of course, they believe their content isn’t providing value to users.”

There is another wider point, however, and that’s looking at the mindset change required for a holistic approach to search.

What does the future of search hold? How should SEO mindsets change?

Search is changing. We haven’t even mentioned the use of platforms such as TikTok as a de facto search or inspiration engine on topics such as food and drink or travel. 

To maintain perspective and plan a holistic strategy, SEOs should remember some key points: 

• • • • • Be aware that search is fragmenting: Google is still key, but entry points include AI assistants, social search, and AI-powered on-site search (think Amazon’s Rufus).

        • Remember, Zero-click search doesn’t mean zero opportunity: Impressions and a strong online presence can be just as valuable. Tracking this impact is key.

        • Think cross-platform: For sustainable success you need to take a cross-platform, conversational and community-powered approach. 

        • Optimise for more than traffic: Start optimising for visibility, share of voice and authoritative content. Spending time and effort chasing clicks alone is no longer enough.

If you’d like to discuss SEO audits or GEO workshops, get in touch.