How do you become part of the AI’s informational context? Your brand’s visibility in the age of AI
SEO Confidential is back – the column that every week brings you the most authoritative voices in SEO and digital marketing, to figure out together where search is heading and how artificial intelligence is rewriting it. A direct, no-nonsense space for anyone who really wants to understand more about the future of the web.
Today’s guest is Giorgio Taverniti, one of Italy’s foremost experts in SEO and digital innovation. Founder, together with the Search On Media Group team, of events such as WMF – one of the largest innovation festivals in Italy – Giorgio is a point of reference for anyone working in online education and communication.
On his YouTube channel FastForward and in his newsletter FastLetter, every week he explains, with clarity and vision, how search engines are changing (or should we say “answer engines”), along with content and digital strategies.
With him we talked about the future of SEO in the AI era, the real meaning of optimization, discoverability, trust, digital culture and why we need to get back to the fundamentals. It’s an interview that digs deep and challenges many of the clichés circulating in the industry.
Take a few minutes and read it carefully: this conversation truly sheds light on where the world of search is heading.
Here you can watch the full interview:
Below, instead, is the full transcript.
“AI hasn’t changed the rules: it is simply highlighting what you should have been doing all along and, in many cases, never did”
Many people keep repeating that “you need a good code structure” or “you have to invest in your brand”, as if it were enough to take cues from an old 2010 SEO manual.
But today, with search increasingly blending with ChatGPT and conversational systems, isn’t it time to stop talking about “optimization” and start discussing how to become part of the informational context that AI builds?
Optimization remains fundamental, but people often don’t really understand what it means. A good code structure is only a small part of the job: real optimization is about the relevance and quality of content, much more than the technical aspects, even though these still retain their value.
The concept of “becoming part of the informational context” is not yet fully understood. Let me give you an example: when you ask an artificial intelligence system for information or a product, the first thing it does is not show you results, but build an internal context.
The AI “tells itself” the question to understand its purpose: why do you want that product? What features does it have? What kind of information is really needed?
At this stage we are not yet in the query fan-out. And if you don’t give it that informational context, the AI creates it on its own.
Now, how do you get inside that informational context? How do you get inside the query fan-out system? And then, how do you get into the product?
These are three crucial steps to map: three pieces of information, three different resources you need to make this transition. Many people have no idea and don’t know what these three resources are. So yes, the informational context is extremely important, and it will become more and more so.
A study by OpenAI and Harvard shows that ChatGPT is used above all to ask for information and advice: in practice, it is becoming the new intermediary between user and content.
So what’s the point of continuing to think rigidly in terms of keywords and meta descriptions, when the game is now being played on the model’s semantic trust and not just on traditional indexing?
Traditional indexing, as of today, has not been superseded. Current search engines work with a hybrid model: they haven’t abandoned the classic system, but they integrate it with the bi-encoder one. In practice, they first use the traditional search engine to identify a set of documents, then apply the bi-encoder system to find the most accurate answer within that context.
Let me give you a simple example: if you search for “smart ring to control your computer”, the system first needs classic indexing to retrieve a minimum set of relevant documents. Only then does the bi-encoder part come into play, working on context and semantic interpretation. The context, at this stage, is provided by the traditional search engine: that’s where everything starts.
At least so far, given how they have been developed. I hypothesized that, in the future, we might move to a system based on embeddings to identify documents directly, but at the moment the big players even seem to be going beyond that direction, so we’ll have to see how it evolves.
But as of today, classic indexing is fundamental, because search engines and AIs need to grab a context. Why do they need it? Because it would cost too much to apply the AIs’ classic system right away. It’s not scalable, it still costs too much. So we haven’t yet reached the model you’re referring to.
By the way, regarding ChatGPT search, between the moment you asked me to do this interview and when I actually did it, I published a dedicated video, because I believe it’s one of the very few valid search systems, perhaps the only one to date.
So, implicit in your question is that it makes no sense anyway to think in terms of keywords and descriptions, because even the indexing system no longer works in that “traditional” way.
And I hope that one day we’ll stop talking about “traditional” and “new”, because in my opinion it no longer makes sense: search engines are search engines, period.
In this video of yours you talk about “industry fatigue” and the slowness with which digital is facing the AI transformation. But don’t you think part of the responsibility lies precisely with those who, over the years, turned SEO into a self-referential technical language, more focused on acronyms than on the basic digital culture that today seems to me at an all-time low?
Look, maybe yes, maybe no. I mean, I didn’t read the questions in advance, because in interviews I much prefer the flow of thought to thinking things through beforehand.
So yes, on the one hand you’re right. In the sense that anyone who has done popularization – and I include myself – has contributed in some way to this problem.
There has always been a bit of self-referentiality, and the language used was often too technical. Sometimes people even overplayed their hand.
I see it today with AI: everyone overplays their hand to pretend they truly master the subject, because a basic culture is missing, and that’s a problem.
But it must also be said that in the 2000s that culture existed. People shared, experimented, and the environments where they did so were forums or personal blogs. There was a more “artisanal” (almost pioneering) level to the discipline, which was very interesting and pushed people to try things, to put themselves out there.
Today that’s no longer the case. There is much more self-referentiality, more personalism: people elbow their way in to grab a few more clients. And the problem also stems from the fact that, for example, a high-school kid who wanted to enter the job market found it easier, until a few years ago, to start with social media than with SEO. At least at the beginning – then every field has its own complexities.
Today, instead, there are AIs, so what happens is that an entire generation approaches digital starting from new tools, but without solid foundations. When you start with social you lack the technical basics, you develop other skills, but not SEO ones. On social there is much more attention to communication, editorial lines, brand – and that’s fine – by the way, I think SEO should do the same thing, when done well.
The problem is that entering the discipline today requires less and less technical knowledge, so people don’t study it. In my opinion, that’s the point. Even though part of the responsibility comes from all that “checklist-style”, superficial popularization which, over the years, has somewhat flattened the whole industry.
You say that “GEO is bulls**t” and that we need to go back to the fundamentals: creating valuable resources and a solid ecosystem. But if SEO today is still smothered by micro-consulting and LinkedIn “fluff”, isn’t that a sign that education too – including that offered by communities and events – has failed to build a truly strategic mindset?
Well no, I don’t agree on this. Micro-consulting, the “LinkedIn fluff”, as you call it, will always exist, but it exists in every industry. We notice it more because we’re SEOs, but if you go into other fields it’s the same. When you bring visibility into a social environment as a metric – both for algorithms and for people – they tend to take things to extremes to get more attention. And on LinkedIn this extremization is evident.
A truly strategic mindset, though, is another matter. I would split it in two: on one side there’s critical thinking, on the other there’s strategy. And in my opinion both are missing today. The problem, at a macro level, is precisely the lack of both. Critical thinking is fading because, at the speed we work, we have too many problems to manage.
I’ve said it about artificial intelligence too: it will go the way of every other technology. Human beings remain what they are, they don’t change. We told ourselves we would work less, but no way: we work more and more, and those who work, work even harder.
That’s why critical thinking is hard to maintain. You always have to stay on top of things, or at least you have that constant feeling, and you end up being more superficial in what you read and in how you reason.
A strategic mindset should instead rise above all this: you should start from the client’s real goals and build a strategy to achieve them. But almost nobody does that. It’s much easier to look at data in the platforms, analyze Search Console or Meta Ads data, than to take the bull by the horns and really work on the direction to follow.
Services are often sold just because people know how to deliver them, not because the client needs them. And the client, almost always, needs something completely different.
I don’t know how much responsibility education bears in this, I think very little. Because here human factors come into play that depend on the individual, not on the educators. You bear real responsibility when you are next to someone every day, not when you give a speech or run an event.
You talk about discoverability as the central philosophy of the future, but in the meantime the market’s attention has shifted towards AI as a production tool more than a thinking tool.
Doesn’t it risk becoming just another slogan if we don’t tackle the underlying economic issue, namely how to redistribute the value that AI is taking away from content creators?
Well, the distribution of economic value is one thing. We talked about it at WMF too: one could imagine a redistribution model, because those who create value must be able to monetize in a different way.
It’s not something that will “happen” easily, but it’s the right thing. Politics should step in, but it’s not simple. The monetization model should be changed, yes, to ensure that those who produce content are rewarded fairly. But I don’t know how this story will end, because I’m seeing many players bypass creators and strike deals directly with the platforms: look at Reddit, look at other similar cases. It’s a problem, a really major problem.
Discoverability, however, is something else. It’s true that the market’s attention has shifted towards AI, without a doubt. And it’s a double opportunity: on one side there are those who use AI to attack and gain ground; on the other there are those who leverage it but need strategy, and here we come back to the earlier point.
Today most companies complain about a lack of strategy. There are few strategic plans, little long-term vision. And that’s because the word “strategy” is not fluff, absolutely not. But it isn’t “let’s do SEO and a bit of email marketing” either. Working on strategy has a completely different flavor.
Discoverability is strategy, because to do it well you need to know clearly who the brand you’re doing it for is, and you need to centralize the information you have, in order to then distribute it the right way. But if you don’t have a strategic plan upstream, you risk making a mess. So, for me, it’s a whole different thing.
As for the underlying economic issue, we can talk about it as much as we want, but it’s like throwing a stone into the ocean: right now nobody is really in a position to change anything. And I see it as very, very tough.
Discover is becoming the single point of access to information, capable of aggregating social media, creators and news outlets into a single stream. But if Google becomes the “homepage of the world”, aren’t we going back to the era of walled portals that the open web was supposed to overcome? How can publishers avoid becoming mere content suppliers for Google’s ecosystem, getting nothing in return?
Publishers must band together, without a doubt. They must ask for what is right to ask, while being aware of what they are asking for. Because we’ve seen it in the past: they have often asked for unsustainable things.
They should be guided by a competent technical committee, capable of formulating serious requests based on concrete criteria. On one side there’s politics, on the other there are the platforms that take and manage the information.
Google is centralizing a great deal, we know that. When there’s this kind of centralization, we’re not talking about innovation, even if right now it seems so. Google does innovate – it has for twenty or thirty years – but this kind of centralization remains risky nonetheless.
Not for Google – for them it’s business, of course – but in general it is an always, always risky situation.
So we need highly competent technical committees, able to truly explain what is happening and to put forward precise, sensible requests. On this I have no doubt.
I believe that, at some point, those who produce content will be paid directly by the platforms that use it. But, as of today, we are still very far from that model.
Discover drives huge volumes of traffic, often invisible in the data. But isn’t it paradoxical that the only channel still truly vital for online news is also the most opaque and uncontrollable one?
Well yes, Google gives us little, very little information about Discover. You can only see it in Search Console and, in part, in Google Analytics as direct traffic. It’s not a good thing, absolutely not.
And it will never really be controllable, because at its core there is an algorithmic feed, so by its very nature it isn’t controllable. There’s no way around this, that’s how it is.
But publishers need to start becoming aware that the environment has changed, and that Discover, just like video, now represents a very important part of visibility.
And yet this isn’t happening: many don’t really take it into account, thinking that opening a “video” section on the site is enough to say they’re dealing with it. But that’s not the case. That part is now one of the most exposed and decisive, and it should be treated as such.
You argue that we should “keep doing what works”, but the problem is that today it is increasingly complex to measure what really works across Overviews, AI Mode, Discover and agentic AIs.
Isn’t there a risk, this way, of thinking only in the short term, neglecting the building of brand authority that can bear fruit only in the long run?
There’s no risk, no. Brand authority is not in question in what I’m saying. When I talk about “continuing to do what works in SEO”, I am necessarily referring to the technical parts. You have to keep doing what already works.
Brand authority concerns awareness campaigns, the brand side of things, and has nothing to do with this dynamic. Those things keep working, and they will continue to do so.
You absolutely must keep working on brand awareness and community, on direct contact with your audience. If you are a brand, authority and community are two things you should always care about: they have always worked.
When I say “keep doing what works”, I mean the more technical parts of SEO. With AI, you simply have to keep doing what already delivers results: if you are producing good resources, keep doing it; if you are doing good work with structured data, keep that up too. These things should be maintained, not upended just because AI has arrived.
AI hasn’t changed the rules: it is simply bringing to light what you should have been doing all along and, in many cases, never did.
If everyone had followed my advice from fifteen years ago – creating an internal Wiki with all their content – today they would be visible in the first phase of AI search’s informational context. But it didn’t happen. Just as, if everyone had kept a blog updated with the best topics and the latest news, today those e-commerce sites would be far more relevant in query fan-outs.
So yes, you have to keep doing what works. And you usually already know what works. What works for AI is something we are still studying, and we will keep doing so. Some things have changed, sure, but mostly in the technical processes.
Entrepreneurs naturally look at sales and conversions; they know very well that traffic for its own sake only serves to inflate egos.
But how can you concretely demonstrate how much SEO really contributes to results? And who can claim credit for a sale, if the user’s journey gets scattered across a thousand invisible steps?
This great storytelling according to which everything on the web is measurable, that “if you spend a cent, you know where that cent ends up” – no way! It’s a nice fairy tale we told ourselves.
In the early years it may even have been true, because there were few channels and very simple attribution models. Today that is no longer the case. It’s a fairy tale, today.
So how can you concretely demonstrate how much SEO contributes to results? There are various attribution models, of course… But when a project involves many channels, each one contributes in a different way.
Honestly, I don’t even see the problem of having to “prove” that SEO contributes. Also because: who does only SEO anymore, today?
It is almost impossible to find someone who does only that. Freelancers no longer do only SEO, because too many different things now impact a project.
So it’s not a question that can be answered easily. The answer, in reality, comes in practice, when you tackle a real project, with a real client. There is no single model that works for everyone.
This is also why consulting today is becoming highly fragmented.
Google is betting more and more on advertising, its main source of revenue. But as soon as you stop investing in Ads, visibility vanishes. SEO, on the other hand, takes time: it serves to build authority and a lasting presence.
The problem is that, if Google keeps hiding users’ data and journeys, it becomes almost impossible to measure the real impact of optimization work. I’m referring to the opacity of Search Console data, for example…
So how can we ask companies to invest in SEO if the tools to demonstrate its real value are less and less precise?
Roberto, this is a very interesting and very timely topic, also because on some projects the value of SEO activities has decreased, while on others it has grown. So it depends a lot on the type of project you are working on.
The SEO’s job today is a job that concerns all visibility on the internet: it is gradually taking over pretty much the entire overall visibility on the internet.
In the e-commerce world, for example, SEO’s impact is expanding enormously across various channels, to the point of working closely with advertising to ensure consistency between the feed and the site.
The truth is that today SEO is becoming mandatory for certain types of sites. If you have an online property, on the other side you find yourself interacting with users, AI agents, search engines, bots and automated systems. It’s inevitable: you have to carry out optimization activities on your project.
In the video I’m preparing for our academy I describe exactly this evolution of SEO. It used to be: Google, users and the website.
Today it has become: internet, users, agents and bots; and no longer just “the site”, but “the digital property”. The scope of our discipline has expanded enormously.
Today we deal with visibility on the internet, no longer just visibility on Google. That’s the truth. And that’s why I don’t feel the need to “prove” the real value of SEO: there’s no need.
We start from the assumption that a company knows it needs visibility. Visibility is either bought or earned, but in both cases it has to be optimized.
As a result, demand for SEO services keeps growing. I was saying one or two years ago that SEO was exploding, and many made fun of me. Now, instead, we are seeing it: there are a huge number of requests for visibility.
So I don’t find myself in the position of having to prove the real value of a consultancy. What we do is simple: the client starts from point A and wants to get to point B. You don’t measure value from a single channel, but from the impact all the work has had in taking the client from A to B.
If you can do that, then you can get into the details. The problem is that, at a macro level, we often still don’t know how to do this. Also because, on the other side, you may be talking with teams that don’t give you the necessary information, and so it becomes difficult to get into this kind of logic.
Personally, though, I don’t see this as a difficulty…
Well, thank you so much for the interview, Roberto, I hope to see you again soon!
SEO today isn’t about being found on Google, but on every corner of the Internet
The future of SEO, then, lies not in nostalgia for “Google’s rules”, but in building authentic value. Those who can interpret the complexity of AI, without losing sight of the fundamentals, will have an enormous advantage.
And this, in my view, is exactly where the real challenge of the coming years will play out: not adapting passively to AI, but engaging in a dialogue with it. Understanding how models build the informational context means learning to “train” the algorithm through solid, coherent and recognizable content.
You shouldn’t think of optimization as a set of technical tricks, but as cultural work on meaning, relevance and trust.
Giorgio says it clearly: today SEO is no longer just about ranking on Google, but about managing a brand’s entire online visibility.
Its field of action has broadened: it involves users, AI agents, bots, search engines and any digital platform where content can emerge.
Visibility today is built and optimized everywhere, because a company that wants to be found can no longer limit itself to Google: it must be recognizable, readable and accessible from every point of the network.
A sincere thank you to Giorgio Taverniti for the depth with which he shared his views on how SEO is changing in the age of AI. See you next week, as always on SEO Confidential, with a new guest. See you soon!
