“Sites that block AI crawlers will lose traffic and be overtaken by their competitors who focused instead on how to show up in LLMs, AI Overviews and AI Mode”, Josh Squires told us
Let’s be honest: the SEO world is in turmoil.
Too often, time gets wasted on anachronistic questions.
Is the old SERP with its reassuring “10 blue links” better, or the new AI Overviews that are rewriting the rules of the game?
Well, that question is a perfect example of a weapon of mass distraction.
The truth is that the future of online search doesn’t wait for the nostalgic.
The classic SERP is destined to disappear, and rather than clinging to what we know, we need to learn to analyze and understand how these AI-generated answers work.
Because it’s no longer just a matter of technology, but of user experience, of new ways of consuming information, and of strategy. The longer you wait to make the leap, the more you risk falling behind.
It’s precisely to bring some clarity, and to put the questions that matter on the table, that I came up with the SEO Confidential column, the space where I interview experts, colleagues and analysts on the most relevant issues in the SEO universe.
For this sixth episode of SEO Confidential I wanted an exceptional guest by my side. An SEO veteran who knows the trenches of digital marketing metre by metre. Ladies and gentlemen, please welcome Joshua Squires, associate director at Amsive.
With over 17 years on the battlefield, Josh is the kind of professional it’s a pleasure to talk to: he has watched trends rise and die, and has had his hands in growth-hungry startups, creative agencies and market giants. SaaS, finance, e-commerce… name an industry, and he has probably left his mark on it.
Josh has been talking about machine learning and artificial intelligence since 2017: that’s exactly why I thought he was the ideal guest for a frank, direct conversation about AI and LLMs.
So, fasten your seatbelt. We’re about to take an unfiltered journey into the post-SERP future. A future where data is hidden and referrers disappear.
Ready to challenge a few deeply rooted stereotypes?
Let’s go.
“It’s not buying and running paid ads that helps, but increasing brand awareness, affinity and interest”: so says Josh Squires
If Search Console doesn’t distinguish between AI Overview and AI Mode data, how can we make informed decisions? And how can we do SEO if Google hides the numbers?
Third-party tools will become the new SERPs. Their measurements will help us understand in detail how our sites are shown in large language models (LLMs), even when answers and layouts are highly personalized and variable.
How often our content appears, its relevance, and the additional information these tools detect, such as images, prices or videos, will give us data similar to what we get today from Google Search Console. If we combine this data with data collected by first-party analytics tools, such as GA4 or Adobe Analytics, we will be able to verify information, make assessments, run tests and develop effective methodologies over time.
We’re not there yet, and it’s a problem that still needs to be addressed, but by combining creative problem-solving, a good knowledge of analytics platforms and the specificity of LLM results, we can build a solid path forward.
Why will some AI assistants still not pass referrers in 2025? Is it really a technical oversight, or a way to hide how much traffic they generate? If Claude, Perplexity and Meta AI can do it, why can’t ChatGPT and Gemini?
Essentially, it all comes down to the encryption and security features of these platforms. Encrypted chats do not transmit referrer data in order to protect user privacy. This is a basic feature in almost every platform that uses encrypted chats, including LLMs and AI chat tools. As marketers, we already faced a similar problem when, in the early 2010s, “dark social” traffic became significant, as social platforms stepped up their security measures. In those cases we relied on advanced settings in web analytics tools and on custom definitions to get rough estimates of that traffic, but not having exact numbers on social traffic never really compromised marketing activities. The same is probably true today for ChatGPT and Gemini.
That said, some referrals coming from ChatGPT are tracked and indicate ChatGPT as the traffic source. So it’s possible that OpenAI has identified situations where referrer data doesn’t reveal sensitive information, or where the query can’t be traced back to the user. For now, this is probably the most marketers can get. Even though these tools are already widely used, we should always remember that they are still under development, with bugs and features that aren’t complete yet. Many things will change over time.
Disabling AI Overview doesn’t mean disappearing from Google: you lose the meta description, not the indexing. Isn’t it time to stop the digital scaremongering and clarify the difference between classic results and AI content? For those working in Search, what is the priority today in the face of the evolution of AI-driven SERPs?
If we’re talking about the impact of AI Overviews and AI Mode, which are progressively replacing the classic “10 blue links” SERP layout, we’re polishing the brass on a sinking ship.
Sooner or later, the traditional SERP will disappear: it’s only a matter of time. What we need now is to deeply understand the nuances of AI-generated answers and start getting used to this new layout and new user experience, not just the technological aspects. Debating whether the old model was better no longer makes sense. It’s going away. We just don’t know exactly when it will happen, and that very uncertainty should push those working in search to adapt even more.
More and more sites are blocking AI bots. After years of free content, should we be grateful that they use it to train their models… without giving us any traffic? Isn’t it time to start charging?
In a perfect world, yes, AI companies would pay a fee to train their models on other companies’ data. Unfortunately, capitalism is increasingly unrestrained, especially in the United States. But most of the training has already been done, and new data relationships have been created to meet upcoming needs. OpenAI’s and Google’s respective partnerships with Reddit are a case in point.
One problem that could arise with a paid model is the same one publishers face when they put content behind a paywall. Paywalled sites get fewer visits, and people often leave faster than on open sites. Economically, though, in some cases this choice can make sense. Sites can assign a value to their data, both for those who read it and for the companies that want to use it to train AI, and use that value to ask for payment. However, looking at how Google behaves with the big newspapers, it seems it would take an agreement among many sites in the same industry to achieve real change. In the meantime, those who block AI crawlers risk losing traffic and being overtaken by those who are working instead to show up well in AI results, such as LLMs, AI Overviews and AI Mode.
Are generic keywords reserved for the big names? Is there still room for newcomers? Does it make sense to invest thousands of dollars if the user is only searching for the brand?
Over the last year I’ve had the great fortune of helping launch a new website in a highly competitive legal industry. In about 8-9 months the business became profitable, without paid advertising, running mainly on organic traffic. Over 90% of that traffic was non-branded. I believe that yes, you can compete for “generic” or non-branded keywords, as long as you pursue keywords aligned with your brand, demonstrate your expertise and authority on the topic, and send the right signals to search engines.
But in this era, it’s no longer just an SEO or search challenge. This challenge requires the full commitment of the marketing department, working in unison toward a set of aligned goals with consistent messaging. I would say emphatically that yes, there is room for newcomers, but they have to be prepared to rise to the challenge.
Does spending on advertising and backlinks help you appear in AI overviews? Or are we just trying to buy visibility in a system that rewards linguistic fame, not technical expertise?
Buying them probably won’t help. At least, not for long. But backlinks generated by honest coverage, business relationships and word of mouth from consumers or your audience do have value and can influence your presence in AIOs. Advertising would work in a similar way: it’s not buying and running paid ads that helps, but increasing brand awareness, affinity and interest. In that sense, your advertising, digital or otherwise, has an even greater need to be targeted at your audience, and your brand must keep the promise your ads make. Success is built on the satisfaction and happiness of your audience.
When it comes to technical expertise, we can’t neglect that aspect either. Everything I’ve just said is true, but it’s also true that we will need technical skills to analyze and identify our brand’s strengths, its measured expertise and authority, and the brands that align with and amplify those qualities. Technical analysis should underpin great creativity and, above all, great products.
Google translates pages into English and shows them in local results, keeping the traffic and conversions. Is this a strategy to “steal” international traffic? And what can be done to defend yourself?
It’s not ideal, I agree. And I’m by no means a defender of Google — believe me, I have a long list of complaints. In this case, however, I believe Google has data showing that there is probably a mismatch between a page’s only available language and the default language of the user’s browser, and that Google is simply trying to serve the user. Unfortunately, this comes at a cost in terms of site traffic.
The other side of the coin, though, is the content itself. If it were valid, effective content, the user might notice the brand and dig deeper. It might answer a difficult question or provide a useful product. The best content will encourage the user to get to know the brand better. If your content doesn’t do that, then the problem is the content, not Google.
If nobody is incentivized to create original content, where will AI find up-to-date information? From social media? From smart devices? And how will brands make themselves known?
Humans will always be incentivized to create original content. Every day there is something new in the world. A new product. A new idea. A new event. There will always be something new for AI to consume. I think this puts pressure on companies to innovate in terms of products, features and value propositions far more than it puts pressure on SEO as an industry or on content creators themselves. For companies capable of doing this, SEO will remain an irreplaceable discipline within the marketing department, and increasingly important in development, as we have seen over the last decade.
In the era of AIO, AI Mode and AI answer engines, what will happen to SEO tools? Will they continue to play an important role in monitoring results, or are they in crisis?
I see most SEO tools adapting. Ahrefs and SEMRush seem to be in a race to ship new features faster than each other, and some really interesting and useful tools are being created. Thanks to AI, I’m also seeing tons of new tools appear. Michael King shared Qforia. Dan Petrovic and the fantastic team at Dejan have launched some really interesting tools.
AlsoAsked has just announced an MCP, which allows their tool to interface with AI agents. There is certainly no shortage of innovative people in this industry, and right now everyone is highly motivated and the challenges are clear. Some tools may fall behind, but conditions are favorable both for the progress of existing tools and for the development of entirely new ones.
Goodbye to the 10 blue links: the challenge is no longer to rank, it’s to become the answer
If you were hoping someone would tell you “relax, in a year everything will go back to the way it was”, you’ve knocked on the wrong door.
The only certainty left is simple: nostalgia is innovation’s worst enemy.
Today’s challenge is no longer just a matter of SEO, but of pure marketing.
Brand and reputation matter, along with the ability to produce content and products so good that they push your audience to seek you out spontaneously.
Randomly bought backlinks and force-fed keywords are just band-aids on a wound that needs serious surgery. You need to build awareness, affinity, genuine interest: that’s where the game is won.
At this point, I know what you’re thinking. You’re worried you won’t be able to keep up. Or that this whole thing has become too technical, too abstract.
And that’s exactly where an experienced consultant can make the difference.
You know, online search has always been an intricate mechanism.
But let me tell you: this new artificial intelligence is not an alien planet — it’s an evolution of those very same foundations.
My job is not to chase the latest trick. It’s to take this new system apart, piece by piece, to understand how it reasons and turn its logic into a concrete strategy for your business. This is where field experience makes the difference, turning theory into tangible results.
Today you’re standing at a crossroads.
You can ignore the change, go on the defensive and hope the storm passes.
Or you can make the strategic choice:
rely on experienced professionals to make your brand so strong and relevant that it becomes the first choice, even for an artificial intelligence.
The game isn’t over. It has just moved to a more interesting field.
The real question is no longer “how can I rank?”.
The real question is: “is my brand strong enough to become the answer?”.
