“Organic rankings have always been a vanity metric and, on their own, they don’t really tell you how much a business is worth”, Despina told us
The world of SEO is going through a profound transformation.
According to Ahrefs data, Google still generates 345 times more traffic than ChatGPT, Gemini and Perplexity combined, yet the sources cited by AI systems rarely match the top positions on Google.
It’s a strong signal: aiming for rankings is no longer enough — today, what really makes the difference is brand authority and content quality.
If for years you have measured the value of your business through the SERPs, it’s time to start looking elsewhere: toward brand reputation, audience trust and the ability to be recognized as a credible source by AI models too.
SEO is evolving into something broader and more complex, where strategy, branding and data intertwine.
To explore all these topics in depth, we interviewed Despina Gavoyannis, SEO and content strategist at Ahrefs and an expert in authentic, human-powered marketing.
With her, we analyzed the new logic of search, the rise of “atomic” content and the increasingly central role of the brand in a universe dominated by AI.
“With less data available, SEO will have to rely more and more on trends and less and less on exact numbers”
Hi Despina, I was reading the Ahrefs data showing that Google generates 345 times more traffic than ChatGPT, Gemini and Perplexity combined. However, the sources cited in AI results rarely match the top positions on Google.
Given these findings, how do you think business owners who have always tied the value of their business to organic rankings should change their mindset?
Organic rankings have always been a vanity metric and, on their own, they don’t really tell you how much a business is worth.
Far more revealing are figures like ROI (return on investment) and ROAS (return on ad spend), which measure marketing effectiveness, while revenue and acquired customers better describe the overall strength and growth of the business.
Even though marketing attribution has become more complex today, the indicators that truly matter are still the ones tied to the real lifeblood of the company: its ability to generate real, lasting value.
Another surprising fact is that brand authority carries more weight than backlinks when it comes to being cited in AI Overviews.
Is SEO effectively becoming a branding battle? How can small and medium-sized businesses compete if visibility seems to be concentrating more and more on big brands?
Yes, branding matters more than ever for SEO today and can directly influence search performance. That doesn’t mean smaller brands are shut out, though: their agility and ability to implement quick solutions can be a competitive advantage over the big players.
Good optimization, especially of on-page elements, can often make up for lower brand authority. In my view, links and online mentions remain fundamental, as they continue to strengthen a brand’s credibility and authority.
AI traffic still seems low in terms of volume, but it has very high conversion rates compared to traditional search. What advice would you give a business owner: aim to maximize these rare but valuable citations, or keep defending classic organic traffic?
Why choose? Pitting the two approaches against each other is pointless.
AI models are evolving rapidly and their adoption is now inevitable: ignoring them in favor of traditional search, which still delivers the best performance today, would be short-sighted.
But the reverse is also true. Even if Google made AI Mode the default for all searches, it would still rank results, especially for systems that use its search API.
Many AI models, in fact, still rely heavily on search engines for information retrieval (RAG) and for queries that require up-to-date results.
Traditional SEO will therefore continue to deliver concrete benefits to brands, allowing them to be represented and showcased within AI systems as well.
Despina, in your article you say that “chunk optimization” is a distraction because chunking is an engineering choice outside the control of SEO professionals.
What does this mean in practice for a business owner who fears losing visibility because of AI-generated answers?
Chunk-style writing is a technique that has been used in structured content since the 1960s and underpins much of the most effective SEO copy of recent years. Business owners who already use it with good results have no reason to change their workflow.
However, when you start obsessing over word counts or the formatting of every single chunk just to improve visibility with AI systems, you end up wasting time.
Worse still, you risk creating content that is less natural and less accessible for real readers.
You talk about “atomic content” as an alternative to the chunking myth: self-sufficient content that is useful both for human readers and when extracted by an LLM.
In your view, what characteristics must “atomic” content have for a brand to genuinely hope to be cited in AI Overviews?
Truly “atomic” content must be designed starting from the outcome you want to achieve: providing a complete, self-contained piece of information that is useful both to a human reader and to an AI model extracting or summarizing it.
Each section should clearly express one central concept, easy to interpret even out of context. It is this approach — more than artificial segmentation into chunks or word counts — that makes a text more understandable, quotable and relevant for AI Overviews.
Many business owners are worried: if AI decides on its own how to segment and use content, how can a brand defend its authority and get “recommended” in the new LLM-based answer engines? Do you think content quality or brand strength will matter more as trust signals in the future?
A brand cannot exert direct control over AI-generated answers, nor actively defend its authority: the algorithms are designed precisely to prevent that kind of interference.
What it can do, instead, is invest in building an authentic brand capable of inspiring trust and affection. When people talk about a brand spontaneously and positively, that authority becomes organic and recognized, and AI systems inevitably end up reflecting it in their answers.
That said, content quality alone won’t be enough. The winning combination will be a solid brand, marketing that truly resonates with people, and authoritative, well-crafted content.
Both traditional search engines and AI systems tend to favor brands perceived as trustworthy and well known, and future technologies will most likely become better and better at identifying and rewarding them.
Ahrefs data shows that 80% of AI citations don’t even appear in Google’s top 100 for the same query.
For a business owner used to measuring success through the SERPs, what does this gap between traditional rankings and visibility in AI answers really mean?
It’s probably the difference between how traditional search works, based on word matching, and how AI-driven search works, which is semantic in nature.
Google had already announced the shift “from strings to things”, and today we are seeing it fully realized. AI search prioritizes understanding meaning, topics and the entities involved, rather than the mere presence of keywords in the text. Traditional SERPs, on the other hand, remain anchored to a lexical model, where literal matching still influences rankings.
The differences in results therefore stem from these two distinct approaches to selecting the content deemed relevant for a query. Over time, the emergence of new models and data will help us better understand how these systems interpret and assess relevance.
AI assistants use processes like query fan-out and reciprocal rank fusion (can you explain what RRF is?), which lead to results that differ from those of search engines.
In your opinion, how should companies change their SEO strategies to tap into these “parallel” logics?
RRF (Reciprocal Rank Fusion) is a method that merges multiple lists of results into a single ranking. In practice, if an AI system queries several sources or search engines, RRF combines the various rankings into one ordered list, giving more weight to results that appear most often in the top positions.
Fan-out queries, on the other hand, are additional questions that the AI automatically generates from the user’s query. They serve to better explore the context of the search, introducing variants, synonyms, intents and characteristics related to the main topic.
For example, from a query like “best running shoes” the AI might generate fan-out queries such as “running shoes for beginners” or “cushioned running shoes” (editor’s note).
That said, companies don’t need to overhaul their SEO strategies. It’s more effective to gradually test new hypotheses on limited sections of the site, evaluating the results before rolling them out at scale.
To tap into these “parallel” AI logics, tools like Ahrefs’ AI Content Helper can offer concrete support: they automatically suggest related queries and subtopics to weave into your content, helping make it more relevant both for traditional search and for AI-based systems.
Perplexity seems more aligned with Google’s results than ChatGPT or Gemini. Do you think we will see a general convergence between AI assistants and traditional SERPs, or is the gap between the two worlds bound to grow?
It’s hard to say for sure. Much will depend on the technical and strategic choices of the development teams, which are still experimenting with different approaches to improve model performance and efficiency.
External factors, such as commercial agreements or investments (for example, Microsoft’s investment in ChatGPT), could also affect the kind of data these systems can access and, consequently, the degree of convergence with traditional search engine results.
The most-cited domains in Google’s AI Mode appear to be Wikipedia, YouTube, Reddit and Google itself. In practice, AI tends to favor UGC (User Generated Content) sources and its own properties.
This leads me to ask: what real opportunities do independent brands have to make progress in this landscape?
The most effective strategy is to target the queries where brands similar to yours appear. In practice, that means focusing on lower-competition keywords, where it’s more realistic to gain visibility in AI answers too. It’s a faster and more sustainable way to stand out than competing on queries dominated by billion-dollar brands.
In parallel, it’s worth strengthening your brand’s presence on the sources that AI cites most often for the most competitive topics, such as forums, communities or UGC platforms.
Even if your site doesn’t appear directly in the results, being mentioned or linked in these contexts can increase your brand’s visibility and credibility both for users and for AI systems.
Today Google offers less and less transparency on the data that really matters to SEO professionals: in Search Console, for example, it’s impossible to distinguish traffic coming from AI Overviews from AI Mode traffic.
In this scenario, how can you make informed decisions? And above all, what’s the point of talking about SEO strategy if there are no metrics to base it on?
It’s getting more and more complex, that’s true. SEO professionals will have to learn to read trends within broader data sets, to spot directions of growth or decline over time. We will no longer be able to count on the same precision in query data or visit attribution we once had, so a new analytical approach will be needed.
It’s still possible, however, to measure overall performance and visibility. Web analytics tools still let you see how many visitors come from each channel and which pages generate the most traffic.
The data will be less precise, but it can still be compared with what was collected in the past. If the metrics show steady growth, it means the site is growing, even if you can’t quantify it exactly. After all, growth remains a positive signal, regardless of the margin of error.
Can you tell us about Brand Radar and how it can be useful in this context?
Ahrefs’ Brand Radar is specifically designed to help SEOs uncover directional feedback when it comes to AI visibility across multiple AI indexes.
The prompts are all based on the most frequent searches users have performed over many years. They are not real prompts.
However, they can reveal patterns such as the number of mentions of your brand over time across millions of queries. You can also spot patterns in how AI talks about your brand, and whether there is incorrect information on the web that you need to fix to protect your brand’s message.
The tool is growing rapidly and will continue to deliver essential features for brands to measure their visibility in AI search.
SEO has become a branding battle
As you’ve seen, SEO is going through a radical transformation: it’s no longer just a technical discipline based on rankings, but an ecosystem that combines reputation, authority and the ability to generate real value.
Artificial intelligence has changed the game, shifting the focus from rankings to relationships, from keyword counts to brand impact.
In this context, a brand’s authority, the quality of its content and the trust it inspires in its audience become decisive factors for visibility in AI systems as well.
As Despina points out, metrics today are less precise and harder to interpret: just think of Search Console, which doesn’t distinguish traffic coming from AI Overviews and AI Mode.
That’s why SEO must evolve, moving from reading numbers to understanding trends, learning to pick up signals of growth or decline over time.
It’s a profound shift: the work of professionals is moving from recording data to increasingly predictive analysis, built on models, projections and intelligent interpretations of user behavior.
Businesses that can read the trends, invest in authentic content and build a consistent, recognizable presence will still have room to grow, even in a web dominated by language models.
A special thank you to Despina for sharing her vision of this new phase of SEO. Her words remind us that growth is no longer measured only in clicks, but in the ability to build trust and value over time.
SEO Confidential will be back next week with a new episode, to keep exploring the future of search and digital marketing together with its protagonists.
