AI search has turned the organic traffic game upside down
Imagine losing 140 million visits in a single year without changing anything about your product, your service, or the quality of what you offer. That is exactly what happened to HubSpot, one of the largest marketing and sales platforms in the world. The culprit was not a strategy mistake or a reputation crisis. It was artificial intelligence quietly changing the way people search for information on the internet.
Today, instead of typing a few words into Google and clicking through links, a growing number of users simply ask an AI and get their answer right there, without ever needing to visit a website. This is transforming not just traffic metrics but the entire logic of digital visibility for businesses. And the central question behind this transformation is one that more and more marketing, content, and technology teams are asking themselves: how do we show up in the answers that large language models deliver to users?
Kipp Bodnar, HubSpot’s chief marketing officer, summed up the scale of the shift well: he used to search the web and scroll through 10, 20, 30 links before finding something useful. Now, people have access to all the intelligence in the world instantly, and the way they find information and make decisions is radically different. This article covers what is happening, with real cases from companies already on the front lines of this change and what they have discovered along the way. 🚀
What actually changed with AI search
For years, the organic traffic playbook was relatively predictable: you produced relevant content, optimized for keywords, earned positions on Google, and received visits. That cycle worked really well for over two decades. But large language models like ChatGPT, Gemini, Claude, and Perplexity have changed user behavior in a way that SEO teams are still trying to fully understand.
The change is not minor. It is structural. When someone asks an AI which CRM is best for small businesses, it does not return ten blue links for the user to choose from. It delivers a direct answer with context, comparisons, and even recommendations. The website that could have been visited simply never appears in the user journey.
This new behavior has a name that is gaining traction in digital marketing discussions: zero-click search. The concept is not new, but AI search has supercharged the phenomenon in a way that no Google algorithm update ever had before. According to Bodnar, the click-through rate for searches that display AI overviews — those AI-generated answers that appear at the top of Google results — is 60% to 70% lower than the rate for traditional searches. That is a brutal number for any company that relies on organic traffic.
The difference from the past is that now it is not just a quick answer box at the top of the page. It is a full conversational experience capable of answering complex questions, making comparisons, and offering summaries that would have previously required visiting multiple different websites. For users, it is incredibly convenient. For companies that depend on organic traffic to generate leads and revenue, it is a red alert flashing on the dashboard.
In HubSpot’s case, the traffic decline came from multiple directions at once. Search engines adjusted their algorithms to fight so-called AI slop — low-quality content generated by AI — which made it even more important for a site to be recognized as an authority on a core topic. At the same time, users themselves started migrating from search engines to AI tools like ChatGPT, creating a parallel stream of information-seeking that simply does not pass through traditional websites.
AEO: the new discipline every marketing team needs to know
Facing this scenario, a discipline that is gaining ground fast has emerged: Answer Engine Optimisation, or AEO. Also called by some Generative Engine Optimisation (GEO), it is designed to help websites appear prominently in the responses of AI tools, including search engine AI overviews and assistants like ChatGPT.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
AEO does not replace traditional SEO. The two disciplines are being used together by many companies. But AEO requires a significant mindset shift. In a traditional Google search, a person types an average of four to six words. In an AI-powered search, according to Bodnar, the average is 40 to 60 words. That is an order of magnitude more in terms of specificity.
This completely changes how content needs to be designed. Bodnar gives a practical example: imagine a company that rents motorhomes in New Zealand. In an AI search, someone might ask for a complete vacation plan for a family of five, including opportunities to see a favorite animal. To be cited in that response, the motorhome company would need to have published something like an article about the most popular animals for kids to see in New Zealand. And that content needs to be written in natural language that matches the kind of questions people actually ask.
HubSpot applied this logic internally and restructured its own content. Previously, the company had long articles explaining its products and how all the features connected. That format lost relevance now that AI can provide those explanations on its own. The new structure uses smaller content blocks that are easy for AI tools to extract. If someone asks about the contact management feature, for example, the AI can locate and use that specific piece of information without having to process a three-thousand-word article.
The results are already showing up. AI now delivers between 7% and 12% of HubSpot’s website visitors in most months. And Bodnar believes this channel will become even more relevant, including in indirect ways: people who discover the brand through LLM responses end up visiting the site directly afterward, inflating direct traffic metrics without the company being able to trace the exact source. As he puts it, people are influenced by LLM responses and then take action through other paths.
Small businesses are getting in the game too
It is not only tech giants feeling and responding to this shift. Spice Kitchen, a company that sells spice kits as gifts, is an example of how smaller businesses can adapt in creative ways.
Ann Lowe, who handles PR and communications for Spice Kitchen, summed up the situation bluntly: to survive, you need to adapt. The company is building what specialists call a content cluster — a dedicated subsection of the website about the history of the spice trade. The goal is to demonstrate authority on the topic so that AI search bots recognize that content as a trustworthy source.
What is most interesting is that this section does not look like a store. According to Lowe, it looks more like a training course. It is content aimed at people who are researching the topic but who end up discovering the brand along the way. It is an attraction strategy that differs from the traditional e-commerce approach, where the focus was always on optimizing the product page to capture the consumer at the exact moment of purchase.
Nathan Pearson, co-founder of Lumos Digital, the agency working with Spice Kitchen, explained this shift in focus: historically, optimization was aimed at the product page, capturing the consumer when they were already ready to buy. Now, the focus is shifting to the research and decision phase, and whoever wins at that point has a huge advantage.
Pearson also recommends that companies publish detailed buying guides. If you have a guide to the best long-distance running shoes, for example, you need to list all the products and indicate a clear winner. AI loves this kind of structure because it makes it easy to extract an objective recommendation to deliver to the user.
Andy Lochtie, also a co-founder of Lumos Digital, adds the importance of expertise, authority, and trust signals. In practice, this includes having many links from other trustworthy sites pointing to yours, linking to high-quality sites from your own pages, and maintaining transparent content policies along with author bios to reinforce credibility. These signals are read by both traditional search engines and language models that scan the web looking for reliable sources.
The MKM Building Supplies case: when AI takes the click but brings the customer
Another case that illustrates this transformation well is MKM Building Supplies, a British building materials chain that sells to both professionals and everyday consumers. Andy Pickup, the company’s digital director, noticed that fewer people were visiting the website because the answers they needed were already being delivered directly by AI models.
A simple example: someone who would have previously visited the MKM blog to learn how to install artificial grass now gets that information in full during a conversation with ChatGPT, without ever needing to open a browser. If that trend continued unchecked, Pickup says, site traffic could shrink to practically nothing.
The company’s response was swift. Pickup recognized it was essential for MKM to be cited in AI model responses. The goal was clear: when people searched for information about building and renovation projects, the language models needed to reference MKM and not its competitors. Beyond strengthening digital presence, this strategy would also have an effect in the physical world, helping drive customers to the stores where the team could assist them in person with their projects.
One data point that really caught the MKM team’s attention: despite Google being the dominant search engine, ChatGPT is sending more visitors to the company’s website than Google’s own built-in AI. Pickup called this a seismic shift in user preference. People are making a conscious choice not to go to Google, even with the search engine’s integrated AI, and are using ChatGPT directly instead.
The strategy Pickup adopted he himself called defensive: creating blog posts about the best-selling products, structured in a way that AI models can easily extract information. He compared the approach to traditional SEO: positioning the company as an expert in certain areas and giving LLMs everything they need to deliver a complete and conclusive answer.
An important difference Pickup noticed is that the content needs to evolve. It is no longer enough to just talk about a product. You need to explain how that product is going to help solve a real problem. Traditional search engines looked for keywords. AI models need to process the meaning of the page with ease. To meet this demand, MKM’s new pages began to include:
- Clear summaries at the top of the content
- Bulleted lists to break information into digestible pieces
- FAQ sections that anticipate user questions
- Updated sitemaps behind the scenes to help AI bots navigate the content
The results speak for themselves. Over the past year, MKM’s traffic from AI tools went from practically zero to a double-digit percentage and keeps climbing. But the most impressive data point might be another one: visitors coming from AI buy more than visitors coming from traditional search engines. Pickup’s theory is that these customers arrive at the site already armed with the information they need, gathered during their conversation with the LLM, which gives them the confidence to complete the purchase.
What LLMs actually value in content
Understanding how large language models choose the information that makes it into their answers is the first step for any strategy that wants to survive in this new landscape. Unlike traditional search engines, which crawl and index pages in real time using well-documented technical criteria, LLMs work from massive training on large volumes of text. They learn language patterns, contexts, associations between concepts, and in the case of models with internet access, they also consult external sources to complement their answers.
This means that a brand’s presence within an AI’s answers does not depend solely on ranking well on Google. It depends on being cited, referenced, and discussed in places these models considered relevant during training or during real-time search. Some patterns have already become clear for those monitoring this behavior closely:
- Content that answers questions directly and objectively tends to be cited more often
- Clear structure with well-defined headings and subheadings makes information extraction easier
- Original data, concrete examples, and unique perspectives increase relevance
- A brand’s presence across multiple channels — such as YouTube, podcasts, newsletters, and specialized forums — boosts the likelihood of association within the models
Another extremely relevant factor is what specialists call topical authority: the idea that a domain or brand that consistently covers a particular topic over time tends to be recognized as a reference by both search engines and language models. This reinforces the importance of having a cohesive content strategy with a clear thematic thread rather than publishing about whatever topic happens to be trending.
SEO is not dead, but it needs company
It is important to be clear: traditional SEO is not dead. It remains fundamental for the vast majority of websites. But it needs to coexist with this new strategic layer that AI search demands. Companies that invest exclusively in one of the two disciplines will be operating with half the strategy the market requires. Those that manage to integrate SEO and AEO organically will have broader and more resilient coverage.
One of the most important points of this new approach is the depth of answers. LLMs tend to favor content that covers a topic more completely, that anticipates related questions, and that presents different perspectives on the same subject. This means a long, well-written, and well-structured article is still valuable, but it needs to go beyond the traditional SEO format, which often prioritized keyword volume at the expense of real information quality. Content needs to be genuinely useful because the models are increasingly capable of distinguishing a text that truly explains something from a text that just looks like it does.
What comes next
The landscape is still taking shape, and nobody has all the answers. What can already be said with confidence is that organic traffic from traditional searches will continue to be competitive, but it will increasingly share space with a new type of visibility: presence within AI answers. Companies that understand this early and start adjusting their content optimization strategy now will have a significant advantage when this transformation fully takes hold.
The cases of HubSpot, Spice Kitchen, and MKM Building Supplies show that there is no one-size-fits-all company for this adaptation. From SaaS giants to small spice sellers, the logic is the same: understand how large language models work, restructure content to make information extraction easier for AIs, and build real authority on specific topics.
It is not about abandoning SEO but expanding it into a broader ecosystem where large language models are one of the main interfaces between brands and users. Business competitiveness in the digital environment has always depended on the ability to adapt quickly. The arrival of AI search is another one of those turning points, perhaps the most significant since Google became the default gateway to the internet.
Companies that treat this as a threat tend to freeze up. Those that treat it as an opportunity to rethink how they communicate, how they distribute their content, and how they build authority over the long term have everything they need to turn this challenge into real competitive advantage. 💡
