SEO for AI: let’s discover the SEO of the future
“In my opinion, search engines like Google or Bing – as we know them today – are already dead and still being used only because there is not yet a high-impact consumer alternative that allows anyone to have access to AI capabilities.” This is one of the introductory, and strongest, passages of “SEO FOR AI. Let’s invent the SEO of the future,” the new book by Ivano Di Biasi, just published with Palladino Editore, with a foreword by Alessio Pomaro.
We publish here a short excerpt from the book, in which our CEO explains precisely his vision on the future of search engines, and therefore of SEO, following the accelerations imposed by Generative Artificial Intelligence systems.
How SEO is changing with AI and how the scene is changing
“Already today, with new models such as GPT-4o, AIs browse the web looking for documents to analyze in real time to provide the answers they did not know.
Current models of AI use search engine results (Bing, in particular) to find the information they need, and there is already internal functionality that interprets our questions, translates them into keywords to search on search engines, performs the search, and processes the results in real time to give us a current and truthful answer.
[…]
There are many players in this very rapid process of change, and each has its own interests to protect.
The search user will have no doubt: he will definitely use a voice or text assistant to search for solutions to any of his problems. Many people already do, and for them there will only be benefits from the new technology.
More than 90 percent of searches happen on smartphones: why open the browser, go to Google, type in keywords, choose a result and read it without knowing whether we will get the answer we are looking for when we could simply ask our smartphone by voice and get an immediate answer?
What’s more, without having to browse websites overcrowded with invasive advertisements everywhere.
AI service providers will certainly have to contend with legal issues, and I believe that every website will have a method to deny AI models from learning from their content, just as Google or Perplexity does with the use of disallow in the robots.txt file, although I see this as a counterproductive choice.
The various AI chatbots are now starting to cite the sources they used to give us the answer, providing a link back to the original article. This is for all intents and purposes search engine behavior, but more modern than Google. At the moment these links are not generated if the AI do not need to browse web pages, but surely, again because of legal implications, they will be far more frequent in the future. It is precisely in this context that I believe the discipline of SEO will shift, which should not take another name such as AIO (Artificial Intelligence Optimization) because it is too generic and not specific to ranking in answers.”
What does SEO for AI mean
“The sense then will no longer be optimizing web pages to come out higher on Google, but optimizing pages to make sure that an AI model can pick our text and link to our website and bring us traffic.
From this consideration, very much against the grain, comes my idea that everyone is going in the wrong direction, which is to try to useAI to do SEO and rank on Google, instead of accepting that search engines will probably disappear along with the SEO we know today and that SEO for AI should be done.
That is, to ensure that we get acceptable traffic to our websites we will have to be able to get AI LLM models to like our content and no longer Google, and everything will be tied to the content we write. It is said that the copywriter’s job will be among the first to disappear-perhaps that is what we are experiencing in this transitional phase-but I am sure that instead the copywriter’s job will be a key role in every company’s visibility strategies.
Websites, too, in my opinion, will benefit from the change as soon as most visits come from AI, and I will explain why.
Websites make money from the advertisements or sales they materialize, little matter where the user comes from. However, if the person comes from an AI chatbot, we will probably never get that visit from Google.
What do I mean by that? I give an example.
If we now Google ‘Yellow squash risotto,’ in Top 10 we don’t find a single result that mentions yellow squash. All the results are focused on pumpkin risotto: Google makes life easier for itself by always responding with the usual well-known/authoritative sites without giving consideration to the user’s need and by bullying smaller sites, which, however, often have the exact answer to our need.
AI models have no bias, are pure statistics and mathematics, and would give visibility in a more democratic way by choosing the best answer to our question.
The point is exactly this: AIs understand our questions better and answer them with exactly what we needed, without dwelling only on what the source is.
If a small blog addresses a particular aspect of any topic that has been overlooked by the big guys it will have a chance of being used as an AI search result. Which is not the case in today’s search engines, which always prefer domains with greater authority, even if the specific page does not meet our need, blindly relying on the site’s authority to avoid looking bad in SERPs.
Everyone will have room to be part of the answers and in this short book I will try to invent, perhaps making some mistakes given the speed at which everything evolves, this new discipline: SEO for AI.
On the advertisers’ side nothing would change: instead of investing in Google Ads they will have to do so in a hypothetical AIAds. They just have to wait for the emergence of advertising business models directly in AI-generated responses, and even then I think it won’t take too long.
The circle is closing around the search engines, or at least for those that will not make deals with AI providers.
When AI will be usable by smartphone assistants and there will be a business model to make publishers and advertisers earn money in search responses, then the current SEO world will be completely revolutionized and only SEO will be done to rank in AI responses.”
How to do SEO for AI: the aspects to optimize on the site
Ivano’s volume thus starts from a set of very current considerations and tries to predict a possible (inevitable?) evolution of LLM models into real answer engines.
In light of this, says Ivano, “it will be absolutely necessary to do a different SEO, oriented to make our websites and our content liked by an artificial intelligence and no longer by the algorithms of a traditional search engine, as well as by users.”
Such new search engines will in fact demand content that is easily understandable to themselves – AI language models – but also useful and relevant to users, regardless of how they do and will do the search.
In concrete terms, this means that if we have detailed content on our websites that meets the various needs of users, we stand a good chance of being chosen by AIs as a source of information. This will not take anything away from classic SEO optimization for search engines, except more attention to turning data and insights about audience intentions into information.
The SEO of the future: goodbye to keywords?
“If we have one certainty, it is that keyword research is a practice that will disappear,“ Ivano writes in his book, carrying forward an insight we had already addressed on these pages as well, summarized in the provocation ‘the keyword does not exist.”
A decade ago, Google’s Hummingbird update was already attempting to initiate a definitive move away from traditional keyword search: users were beginning to use increasingly precise phrases in their queries, hoping to get detailed answers on specific topics. It was an early turning point, as it aimed to better understand and answer complex, conversational queries by reducing the emphasis on exact keywords.
The trend toward more complex searches was fueled by the rise of voice searches and an increasing need for precise answers. Search engines had to evolve to correctly interpret the context and intent of queries, using technologies such as Natural Language Processing (NLP) and semantic indexing (LSI) to understand synonyms and related terms.
Today, queries and FAQs are becoming the new fundamental units by which searches are conducted, both on search engines and on AI platforms such as ChatGPT. This means that, in doing SEO for AI, the concept of keywords must be replaced by that of FAQs. It will be necessary to measure the search volume of questions related to a topic rather than individual keywords, recognizing that everyone formulates questions and searches differently.
The use of AI in SEO promises to solve many of the difficulties of interpreting varied queries, ensuring accurate answers even when the question comes from an inexperienced person. However, it seems that Google has reduced the effectiveness of technologies such as Hummingbird, with less accurate search results. Therefore, in terms of SEO, the focus should shift from ranking by long-tail keywords to building brand or website authority through link building strategies-not least because, as seen with Google’s recent behavior study, the long tail is about to disappear prerogative of the big sites that monopolize the SERPs, capturing the famous 45.46% on all search intent keywords.
What tools to use to do SEO for AI
Let’s go back to the practical and concrete aspects again: to adapt to the artificial intelligence-based SEO of the future, it is essential to use tools that allow us to better understand users’ search intent. Even before the announcement of Google’s AI, Gemini, it was evident that FAQs were becoming a crucial element in SERPs thanks to the “People also ask” box. This indicated that for each keyword there are a myriad of related questions, each with a different search intent. Google, in fact, seems to anticipate users’ needs by proposing related queries, as understanding exactly the intent behind a single keyword can be complex.
To address this challenge, SEOZoom has introduced an indispensable tool: the Question Explorer. This tool, developed with the goal of collecting and analyzing all questions that emerge from user queries, represents a further step forward in keyword research. Indeed, the Question Explorer makes it possible to build an almost infinite database of FAQs, related to a wide range of keywords, revealing the search volume of the questions themselves and the potential traffic they can generate.
Thanks to this tool, we can conduct keyword research no longer based solely on keywords, but on the real needs of users. Moreover, it allows us to identify the most frequently asked questions in our industry, foreshadowing future scenarios in which we might interact with AI search engines via voice commands from the smartphone.
In summary, SEOZoom’s Question Explorer is currently the only tool in the world that allows keyword research based on user queries, making it indispensable for both traditional SEO and future artificial intelligence-centered SEO. Using this tool allows for a better understanding of users’ real intentions, adapting SEO strategy to the evolving needs of the digital landscape and increasing the chances of visibility: no longer on a search engine, but among the answers chosen and provided by AI.
In short: again, SEO is not dead, but it needs a defibrillator. And SEO for AI offers vital insights, tests and experiments for those who work online and want to anticipate the changes needed for tomorrow.
Excerpt from “SEO for AI. Let’s invent the SEO of the future” by Ivano Di Biasi, Palladino Editions ©2024, reproduced with permission of the author. For information: SEO for AI.
