Hi, I’m Anton Sednin, Campaign Manager at Grid Concepts, and I want to talk about how GEO (Generative Engine Optimization) and Google AI Overviews have changed the way our marketing campaigns work in 2025-2026. All my conclusions are based on statistics from our clients in Melbourne, Brisbane, and Australia as a whole. Only real data, only hardcore!

This content may sound a bit professional, so I’ll define the audience right away. This article will be of interest to you if:

• You’re also an SEO specialist, and you’re interested in how competitors like us have adapted to new AI trends
• If you’re an SME owner in Australia and you’re interested in how other small business owners’ websites are doing
• If you’re an SME owner and you’re currently at a crossroads and wondering whether to spend additional budget on GEO marketing (spoiler alert: it’s not a good idea).

So, the basic questions are:

• Should you spend money on GEO? (Spoiler: no)
• Should you spend money on AEO? (Spoiler: yes)
• How to retain leads in 2026 and 2027?

Our insights on all the questions are below

Let’s go.

Is AI a good source of leads and traffic for local businesses in 2026?

In general, the answer is ‘no’.

But let’s dive into the details.

Many business owners in Australia experienced a sharp drop in traffic and leads in July-August 2025 (sometimes even dramatically, as we’ll discuss below using one of our clients as an example). Meanwhile, all the talk around is ‘Hey, AI rocks! It’s the new global trend!‘. As a result, it’s tempting to spend money on GEO (in simple terms, a website optimization for neural networks sources of traffic).

So why not spend a little money to appear in ChatGPT?

And again, no.

Our team works with dozens of small and medium-sized business owners in Australia with traffic ranging from 3,000 to 500,000, and if you came to us for a consultation right now with this question, we’d likely tell you that GEO isn’t worth it. We received very few leads from neural networks in 2025-2026. Using AI visibility testing services like keyword.com as a benchmark, we have local Australian business websites that have strong visibility in ChatGPT, but only 2 out of 100 monthly leads come from neural networks.

Here’s a screenshot of the traffic share from neural networks for one of our clients:

Pretty low, right? And this is one of the clients with good visibility in ChatGPT.

But ChatGPT isn’t the only neural network. So, what other neural networks drive traffic for our clients? Here are the top ones:

• chat gpt
• perplexity
• claude.ai
• copilot.com

Yet ChatGPT brings in more traffic than all the others combined. Our largest client had 500,000 views last month, of which ChatGPT accounted for 1,500 (0.3%). Incredibly low.

And no, it’s not because we’re bad at SEO (although how will you check this? Ha-ha), but simply because most SME owners’ websites are not well-suited for GEO.

So, should local business owners in Australia consider GEO?

More likely yes than no. The hype in this case is overrated. The paradox is that if you ask ChatGPT right now, the AI will answer something like “‘Good idea, buddy! In 2026, GEO is important specifically for local businesses, go ahead!‘

So, what’s the contradiction here?

When displaying information about local businesses, neural networks rely on reviews and Google Maps. For example, if you provide cleaning services in Melbourne and have the best reviews on Google Maps in your area, then congratulations, there’s a 99% chance you’ll be mentioned by ChatGPT.

In other words, many business owners need to take care of their Google Maps profile and forget about GEO in general.

Here’s the key thing local business owners need to understand: the trend toward focusing on location began long before GEO appeared, and Google has been emphasizing the importance of local traffic for two years now.

Here are some statistics from our clients that illustrate this point.

Client Example #1: It shows that traffic spikes from Australia began in August 2025:

Client Example #2: Here we see that the site fell into recession in 2025, but traffic from Brisbane (the client’s primary location) remained, and even showed growth in 2025:

We’ve observed this trend for almost all of our Australian clients since late 2024. In the screenshot, you can see growth spurts in August and September (It was just at this time that AI Overview was introduced). After a seasonal decline in December, the trend continued throughout 2026.

Therefore, in 2026, the correct answer is: local businesses should prioritize demonstrating their local presence online more seriously, and only then consider their GEO.

It’s important to understand the difference between a local trend and a GEO trend, as this ultimately determines promotion methods.

So, is GEO useless for local businesses at all?

Not quite. In fact, the real race right now isn’t to get into neural networks like ChatGPT, but to get into AI Overviews (Google-generated answers that appear above the very top search results). Оver the past six months, SEO specialists have realized that it’s in AI Overviews that the real battle for traffic is unfolding.

This is why GEO recently spun off into a new division, AEO (Answer Engine Optimization), which focuses specifically on AI Overviews and Featured Snippets.

We see from our clients’ responses and Google search statistics (see screenshot above) that small businesses in Australia still have a poor understanding of what AEO is in 2026, but 9 out of 10 predict that by 2027, every SME owner in Australia will know this acronym.

Which local business owners in Australia have lost traffic after implementing AI Overviews?

If you’ve been following this trend, you’re generally aware that AI Overviews have primarily impacted news, lifestyle blogs, how-to guides, and so on. Basically, any type of content that’s easy to skim without requiring a click.

Pretty simple, right?

But what about the websites of local businesses in Australia? It’s a bit more complicated.

Here’s a screenshot showing an example of the most dramatic decline for one of our clients in July 2025:

In one week, traffic dropped by 90%. Blog content essentially ceased to exist. I won’t go into the full list of measures we took. I’ll just share the main findings about the reasons for the drop:

• The website was old and had a poor mobile version.
• The author wasn’t listed on the website.
• The website had poor-quality blogging (no interlinking, no ‘recent posts’ section, etc.).
• The blog section was on a subdomain.

Simply put, it was an old website, built back in the 2010s, and the client had been delaying updating it. Similar situations were frequently reported on forums in August-September 2025. Even our clients who maintained TOP-3 ranking pages couldn’t break into AI Overviews without an author and poor-quality blogging.

Simply put, Google not only implemented AI Overviews but also tightened its SEO criteria. More precisely, Google made the trends of the last 5-10 years more important. Specifically, we’re talking about trends like:

• Mobile First
• Featured Snippets
• EEAT

Since August-September 2025, these features have become exponentially more important.

For example, things like authorship transparency in blog content creation are called EEAT. Simply put, it’s proof of the content creator’s authority and expertise. This is largely what destroyed 90% of the website in the example above. The EEAT concept has existed since 2022, but since August 2025, its importance has increased dramatically. Australian clients with high-quality EEATs were unaffected, while clients with poor EEATs saw their rankings drop.

EEAT sounds simple, but in reality, EEAT encompasses a variety of things, such as:

• ‘About Us’ and ‘Terms & Conditions’ sections on the website
• Indication of a real physical address on the website
• Transparency of expertise (real photos + real author of blog posts with a social media profile)
• Public profile of the company owner
• Reviews on Google Maps
• High-quality backlink list
• At least 70 out of 100 on pagespeed.web.dev for both desktop and mobile versions.

In my eight years of experience in SEO, I’ve seen about 30 websites (mostly Australian) and have never seen all of these points met. If you think you’re generating quality content but are still losing traffic right now, just review the list above. You’ve almost certainly missed something.

Which local business owners in Australia saw their traffic increase after implementing AI Overviews?

All of our clients with good service pages saw their traffic continue to grow, and keywords improved dramatically. I gave an example above of a client who lost 90% of their traffic on a subdomain with blog content. And here are the same client’s statistics for the same period on their main domain:

This main domain is essentially a list of service pages. The screenshot shows that traffic has remained unchanged, while keywords increased sharply in September 2025.

Many business owners in Australia believe that the drop in traffic has affected almost all websites, but this is not the case. Here is a video featuring GA4 and Search Console statistics for two of our clients, clearly demonstrating that the introduction of AI Overviews actually improved the traffic:

This is a common trend for all of our clients. And it’s another signal for business owners: it’s time to review your service pages and improve them as much as possible.

What works for AEO when promoting local businesses?

Now let’s dive deeper into how AI-powered SEO works.

To make things clearer, we’ll break this section down into three steps. Each step is important, but each subsequent step is more complex and deviates further from standard SEO to specific AEO.

Step 1: Follow basic SEO because it also works for AEO:

• Write meta tags
• Check that Google isn’t overriding your meta tags (if Google thinks a meta tag doesn’t match the page, it changes it)
• Meta tags must be 100% unique
• Content must be 100% unique
• Check your breadcrumbs.
• pagespeed.web.dev should show at least 60+70 out of 100 for both mobile and desktop.
• Search Console should be error-free.
• Sitemap and robots.txt should also be fine.

Step 2: Use basic AEO

• Add FAQ sections. This has become more important in 2026.
• Intent analysis: try to imagine what questions the user will ask to AI and answer them in the FAQ section.
• Find all Featured Snippets on the topic and also answer them in the FAQ section.
• Create additional markup for the FAQ section (AI loves it).
• The first two sentences in FAQs should be a comprehensive answer to the question.
• Spend more time on internal linking. Even if you have 5 articles on your site, they most likely contain terms that carry over from article to article. Just imagine you’re a Wikipedia editor and interlink all the terms and services as if it were an encyclopedia article.
• Make sure the ‘dateModified’ parameter in your schema markup is updated. Simply put, the page needs to be up-to-date.

Step 3: Advanced AEO

For more serious AEO promotion, cross-linking is usually necessary.

For example, if you sell robotic vacuum cleaners in Brisbane and post about your robotic vacuum cleaner tube on Reddit and Quora, you’re more likely to be noticed by ChatGPT and AI Overviews.

In simple terms, advanced AEO is working with link authority.

And if you’re an SEO specialist, I know what you’re thinking: why make things so complicated? Why not just buy a few backlinks? But AI doesn’t like backlinks as it is; they like reliable sources.

And this also answers the question of how backlink strategy has changed in 2025 compared to 2026: AI are more willing to use data from Wikipedia, Quora, Reddit, rating aggregators, Google Maps, and review sites. And the more interested a website owner is in GEO and AEO, the more attention they pay to it.

Our team’s experience shows that you’re more likely to get into AI results if you have 10+ reviews on Google Maps than if you’ve purchased 10+ links with high domain authority. And this is a clear signal to business owners that in 2026, it is better to spend resources on Google Maps reviews than on buying backlinks.

What types of articles perform better after implementing AI Overviews?

The above guide is of little relevance if you have an informational website, so it’s worth discussing separately. There are different types of blog posts, for example, guides, interactive content, evergreens, etc. What has changed in this regard since implementing AI Overviews?

Almost all Evergreen articles stopped working:

• Top 10 articles
• How-to articles
• Ultimate Guide articles

The following performed well:

• React Content (an expert reacts to a news item relevant to their business)
• Living Content (for example, articles with real estate interest rate forecasts. This topic can be updated 1-2 times a month due to constant news)
• Cornerstone Content (the main fundamental article around which the cluster is built)
• Freshness-Driven Content (focus on regular updates for the sake of freshness.)

Here are the statistics for the distribution of content types on the Australian information portal of our client:

These data clearly show that Living content has become significantly more popular. However, this data doesn’t reflect the fact that we now have to expend significantly more effort to ensure that Evergreen content maintains its positions.

Here are some more statistics. This is what the decline of a typical TOP-10 article looks like after August 2025:

And here’s what the Living Content article looks like for the same client after July 2025:

As the screenshots show, the shift in blog posting trends occurred quite suddenly in August-September 2025.

How have traffic sources on Australian websites changed since the implementation of AI Overviews?

It’s not just the type of articles that changed, but also the sources of traffic for them. So, let’s summarize these changes:

• Traffic has deglobalized
• The role of mobile traffic has grown
• The role of desktop traffic has declined (once again, ha-ha)
• Traffic has become less keyword-driven and more source-driven (a well-known characteristic of Discovery traffic)
• The role of news traffic has increased (apparently, for the same reason: a good news site is always more about platform authority and less about keywords)

Let’s look at the situation using one of our clients as an example. Here’s the traffic distribution for January-May 2025 and January-May 2026:

Yes, we lost about 20% of Google Search (what a surprise!), but what’s interesting is that the share of local traffic dropped by only 13%, and almost all of this decline came from one of Australia’s three main cities. In other words, we’re getting the same number of users from almost every city in Australia as before except one (sorry, we won’t say which one, otherwise you might be our competitor, ha-ha).

It’s not as scary as many SEOs wrote last year, right?

This is something we’ve already discussed above, using examples from our other clients: the traffic decline that was widely reported in 2025 and 2026 due to the AI trend is primarily due to eliminating irrelevant user traffic and tightening criteria for local presence.

Simply put, we now have fewer Indian and Chinese visitors, and our traffic is showing up less in cities we don’t cover.

Here’s another big thought: the negative impact of AI Overviews is primarily related to desktop traffic. It’s obvious, but analytics don’t often mention it. It’s on desktop that AI Overviews perform best.

Meanwhile, Google Discover’s share (when the user isn’t intentionally searching for information, but receives a recommendation from Google) increased by 30%.

As the following screenshot shows, Google Discover traffic comes in bursts:

Although the graphs above don’t reflect this well, compared to 2025, Google News’ share of traffic has grown significantly.  

Overall, traffic has become more tied to current events, and competition for trending content, freshness, and relevance has increased. SEO has become more like a raging river, where the winner is the one who acts not only technically competently, but also quickly and boldly. 

O tempora, o mores! 

SEO vs. GEO. Which should you choose for website promotion?

Based on all of the above, the short answer is: SEO is better. Choosing GEO only makes sense if AI sources are bringing you at least some leads, and you have positive dynamics for this type of traffic in GA4. 

Our clients’ statistics show that AI doesn’t bring any leads in 9 out of 10 cases (yet?). 

Is it worth installing llms.txt on a website in 2026?

I wouldn’t rush into it. 

I haven’t met anyone who installed llms.txt on their website and saw any positive changes. Neither OpenAI nor Google have confirmed support for this feature yet. It’s not a given that this feature will even work. At the same time, our clients’ statistics show that in a year or two, something will definitely need to be done about the abundance of AI bots on websites. 

Here’s an example. In December 2025, our team was shocked when the share of bot traffic for several clients increased from 10% (the default for websites) to 50%. And these were all AI crawlers. There were so many of them that it seemed like a DDoS attack. Moreover, the problem turned out to be widespread. By early 2026, everyone in the SEO industry was talking about the increase in bot traffic.

Here is a video featuring GA4 and Search Console data for two of our clients, clearly demonstrating the correlation between the onset of a sharp surge in bot traffic and a subsequent drop in impressions in SC:

 

If you’re an SEO specialist, you understand how shocked we were when we saw this data. 

Both sites have 5-7k monthly traffic, one serves a Brisbane audience, the other a Melbourne one, and both began suffering from massive bot traffic from China at almost the same time. 

What does this mean? The overall idea of llms.txt is to create clear rules of the game between AI and site owners. And the screenshots above illustrate that such rules don’t exist yet. Currently, site owners are faced with unpleasant surprises like the example above. 

Nobody likes surprises. Everyone wants to play by clear rules. Does llms.txt help make the game more predictable now? No. Will it help in the future? Oh, we all hope so. 

GEO requires short answers to questions. Isn’t that the same as Featured Snippets? 

All AI guides say that for GEO optimization, you should try to imagine what questions users will ask when interacting with a neural network. But at the same time, tailoring text to short and clear answers is what SEO specialists were doing 5-10 years ago with Featured Snippets. Aren’t they the same thing? 

Indeed, Featured Snippets and GEO overlap significantly here.

But there is a clear difference.

When a user asks ‘How to fix a vacuum cleaner?‘, this question is definitely a Featured Snippet. But what if the user asks, ‘Why does my dog keep chasing the robot vacuum cleaner?’ That’s definitely NOT a Featured Snippet. But it could be useful for those selling vacuum cleaners. See the difference? 

Here’s a more abstract, but perhaps more understandable example. 

If you are a SEO-specialist, you’ve probably encountered a situation where the query is so narrow that the only links it yields are to forums like Reddit. But forum links are always a problem because you can’t use tools like Ahrefs to parse the keywords from there. So, there are no keywords, but there is a link. The difference between GEO queries and Featured Snippets is very similar to the difference between the questions you see in Google and the questions you see on forums. 

By the way, if you’ve ever done copywriting, you know very well that forums are an excellent source of information when it comes to creating added meaning. It’s a deeper layer of the Internet, so to speak. It’s the Internet as it was from the very beginning, back in the 90s. 

Essentially, neural networks are trying to capture this layer of information. 

Should you avoid AI Overview-based keywords when generating content? 

Today, the answer is obvious: no, you need to check the AI Overviews and try to write your content in such a way that Google will add you as an additional data source to the AI Overview. 

But this question has become one of the most discussed topics in our team in 2025 when it comes to blogging. 

Imagine the situation: you’re preparing some content plan, but the keywords you’re proposing are unable to rise above the top 10 due to the domain authority gap between the client and competitors. This wasn’t a problem before, as a top 15 ranking wouldn’t be bad either. But today, if you’re not in the top 10, you’re not in AI Overviews. As a result, every SEO specialist wonders: wouldn’t it be easier to write blogs on more specific topics for which the AI Overview isn’t available? 

One solution we found was to update the information. Abandon the evergreen strategy and strive to answer the most pressing questions in our clients’ business niche (which we discussed in more detail above). 

Is it true that AI Overviews ignores pages that rank in the TOP-1?

If you’re an SEO specialist, you might have noticed that some of your articles rank in the top 3, but AI Overviews still cites articles from lower rankings. It seems Google did this deliberately to compensate for the decline in the number one search page, which was widely reported in 2025. 

We’ve also encountered this with our clients. But I wouldn’t build a conspiracy theory around it. It’s just that good structure is important for SEO, and authority is important for GEO. If you have perfect content, but a competitor has higher authority (see above about EEAT), a paradox arises where a page ranks in the top 1, but AI Overviews cites those in the TOP-10. 

Is it true that updated content ranks better in AI Overviews than older content? 

No, it’s just like standard SEO: if you created a useful page and the information in it is still relevant even five years later, you won’t lose rankings because Google prefers older content over newer content. 

The reason people are asking this question today is that many websites experienced a dramatic decline after the introduction of AI Overviews. But if your site has been technically sound and has followed SEO trends for the past 10 years, older posts should rank better in AI Overviews than newer ones. 

How much will adding an FAQ increase my page’s chances of being included in AI Overviews? 

I hear everywhere that FAQs play a huge role in GEO, but I’d say this feature isn’t that important. 

I have two examples of local clients. Both have good reviews on Google Maps. Both have services pages with a TOP-10 ranking. But the client without an FAQ section was included in AI Overviews, while the one with an FAQ wasn’t. 

The problem is that the first competitor operates in a niche with only a dozen competitors, while the second one operates in a niche with hundreds. 

I’d say that FAQs have become more important than before. But they’re no more important than a link profile, Google Maps ranking, etc. 

Does using neural networks help our team boost sales to our clients? 

In 2020, texts were written by hand, but in 2026, they’re generated by neural networks. In 2020, designers created separate images for Google Ads and Facebook Ads, but now it’s relatively easy to create videos. Has this helped us increase profits for our clients? Rather, it’s helped us maintain our metrics at the desired level. This is especially true for video content production. 

Oh, we produce a ton of video content compared to 2025. 

Overall, using neural networks isn’t some kind of cheat code, but rather a matter of trend: everyone uses neural networks these days, so if you want to maintain things as they are, you should too. 

But paradoxically, our team has plenty of cases where a dozen generated videos, after a $3-5,000 budget in Google Ads, generated fewer leads than traditional images. But there are also counterexamples, where we manually rewrote the text several times, and a good prompt pushed the article into the top 5. 

Have things become easier? Oh, yes. But the freed-up time has to be spent on higher-quality SEO to achieve the same results as when we joined 3-4 years ago. 

What does this mean? Our expertise is now more in demand. While previously a client could say, ‘I’ll write this article myself and get traffic’, and they actually succeeded, things have changed. The client may not even write the article (ChatGPT will do it for them), but without proper SEO, the article won’t get traffic at all. 

At the same time, the examples above show that the AI trend has brought a lot of chaos to SEO. This makes clients (and, of course, Australian interest rates, ha-ha) more wary.