Google has thrown another grenade into Search, extending the AI features that are already making life harder for publishers, website owners, and marketers. Announced as part of its annual I/O developer conference, the new changes revolve around transforming the classic Search box into an “intelligent” window, one which expands according to the user’s queries and which can process multi-step prompts. Google has also integrated its now-familiar AI Overviews feature with AI Mode, which users can seamlessly use to input further queries after receiving an AI-generated summary.
The Big Tech giant is calling this “the biggest upgrade to our Search box in over 25 years”. In addition, it’s rolling out Search agents, which ”operate in the background 24/7” and provide regular updates on and summaries of topics the user has asked to track. The idea here is that the user will have to perform fewer manual searches themselves, and while it does look like a potential timesaver, it’s already beginning to unnerve SEO teams the world over.
Traditional optimisation models are rapidly becoming redundant
Indeed, several SEO experts have told Raconteur that traditional optimisation models are rapidly becoming redundant. But while they agree that Google’s update will likely increase the amount of time users spend with AI, they also affirm there are many things publishers and website owners can do to adapt. This includes prioritising original, first-hand information and content, as well as strengthening E-E-A-T (experience, expertise, authoritativeness, and trustworthiness) signals that increase the probability of being cited by AI. So even if it may now be impossible to reverse the move towards more AI in search, the most proactive SEO and marketing departments will be able to carve spaces for themselves within Google’s brave new world.
AI Overviews already eating into search
Probably the key aspect of Google’s Search update is that it’s not simply about introducing new technical changes. It’s also about encouraging and cultivating behavioural changes on the part of users, whom Google is now inviting to spend more time within the confines of its now-intelligent Search box. This is also reflected in the inclusion of an AI Mode query box at the bottom of AI Overviews, so that instead of clicking on a traditional search result, users can continue having a conversation with AI.
For SEO professionals, these changes will accelerate a process that began in 2024 with the introduction of AI Overviews. This is the view, for example, of Rankstar CEO and founder Thomas Kraska, who tells Raconteur that a 2025 study from Pew Research Center has already demonstrated a significant drop in organic clicks following the initial rollout of AI Overviews.
“Yes, the evidence is solid and it’s only going to get more pronounced,” he says. “Pew’s 2025 study already showed clicks on traditional links dropping from 15% to 8% when an AI Overview appears, and users are 26% more likely to just end their session right there”.
The first-placed Google search result suffers a 58% dip, on average, when an AI Overview is present
Kraska also cites data from Ahrefs, which indicates that the click-through rate for the first-placed Google search result suffers a 58% dip, on average, when an AI Overview is present. In fact, this was the rate of decline in December, while back in April 2025 the percentage stood at only 34.5%. The current percentage could therefore be significantly higher.
“The new I/O updates — Gemini 3.5 Flash powering a smarter search box that flows straight into AI Mode and agents — will push even more users to stay inside Google’s interface for conversational or research-style queries”, he adds. “For informational content especially, people are getting the answer without leaving the page”.
Other SEO specialists have a similar outlook. “Google’s new AI-integrated search box and search agents will undoubtedly lead to fewer clicks”, says Brian Dean, the founder of Exploding Topics. Dean also notes that AI Overviews have already dented search clicks, yet he adds that it’s too early to predict the long-term impact on website owners, particularly when other negatively perceived changes — such as the introduction and evolution of mobile SERPs — ended up increasing total clicks overall.
Don’t let Google E-E-A-T your website’s lunch
Likewise, Sam Robson — the CEO and co-founder of the better web co. — agrees that the long-term trend is towards AI becoming an interface for the entire Internet. Yet he also warns that the precise impact of Google’s latest update isn’t yet clear, particularly when AI Overviews have already had a sizeable impact on clicks.
“However, they’ve been so impactful, there’s a genuine question of whether these new changes can actually make things worse for information content”, he tells Raconteur. “I’ve seen some sites who already have 90% of their SEO keywords serve an AIO [artificial intelligence optimisation], so there’s very little left to take away”.
Regardless of whether the latest update may be the decisive factor, every SEO expert Raconteur has spoken to agrees on the general prognosis. “It’s not total doom, but the old ‘top 10 links win’ model is eroding fast”, says Kraska, who advises every website owner to begin evolving their SEO practices, assuming they aren’t already.
“The core game is no longer just ranking #1 for blue links — it’s becoming a trusted source that Google cites inside the Overviews (and now the new AI agents)”, he adds. Accordingly, he recommends that website owners focus increasingly on E-E-A-T signals. This means placing greater importance on original data and research, providing clear and concise answers, using strong schema markup, and publishing content that’s easier for AI to process and quote.
For Kelly Cutler, an associate professor in digital marketing at Northwestern University, the ongoing shift towards more AI in Google Search requires a “fundamental” evolution in SEO and digital marketing practices more generally. She says, “Traditional SEO must expand to include generative engine optimisation (GEO), with a focus on visibility and measurement within zero-click environments.”
Similarly to Kraska, Cutler advocates systematic use of structured data and schema markup, optimising for AI-based citations, and building authority through the harnessing of community and trusted platforms. “Tools such as SEMrush and Ahrefs, along with emerging AI visibility frameworks, will play an increasingly important role”, she explains. “At the same time, platforms like Google Search Console and Google Analytics 4 remain essential for benchmarking performance and understanding shifts in traffic sources and user behaviour”.
99% of AI citations or links are from non-paid media, while 82% come from earned media
Again and again, SEO experts argue that original content will, perhaps counterintuitively, have much more value in a scenario where people are predominantly using AI-based search tools. This is borne out by recent data from Muck Rack, which found that 99% of AI citations or links are from non-paid media, while 82% come from earned media. More granularly, 25.4% of links are to journalistic content, 24% to corporate blogs and content pages, 16.7% to encyclopaedic and reference sites, and 15.2% to owned media. This indicates that Google and its users don’t want AI slop to rank highly, something affirmed by Brian Dean.
“This is why unique data, expert perspectives and opinion pieces are performing best right now in SEO and GEO”, he says. “But to get the clicks that are left, publishers need to put out unique stuff that ChatGPT can’t write”.
The importance of autosuggestion
There are other strategies for ensuring AI visibility, with Thomas Kraska revealing that Rankstar is now helping its clients optimise for Google Autosuggest, so that a brand name or webpage will appear before a user finishes typing a search query. “It puts you in front of users at the earliest possible moment and drives traffic that bypasses the AI Overview layer entirely”, he says.
Kelly Cutler highlights another increasingly important area of discovery: social networks and user-generated content, with sites such as Reddit counting among the most frequently cited by AI models. “Platforms like LinkedIn, YouTube, and Reddit are central to discovery, credibility, and community building”, she explains. This implies that, once again, brands and businesses must focus on cultivating authority, on providing the kind of original content and value that increases the probability of mentions.
In the slightly more distant future, Cutler also predicts that advertising will follow attention, implying that paid advertising will increasingly insert itself into Google’s AI Overviews and AI Mode. She notes that this has already happened with ChatGPT, and that continued innovation from Google is likely to create an expanding space for increased monetisation.
Brands and businesses must focus on the kind of original content and value that increases the probability of mentions
“New Google products like Gemini Spark (AI personal agent), Gemini Omni Flash (video creation tool), and SynthID (AI content tool) from Google will help marketers deliver quality campaigns at scale’, she says. “The competitive landscape is intensifying, particularly as companies like Meta invest heavily in AI-driven ad creation and delivery”.
Given that current AI models cite only a small handful of sources, competition will also intensify among the websites aiming to appear in results. And when you add the fact that fewer people will evidently be clicking on traditional search links, businesses will have to become very proactive very soon.
As Cutler concludes, “Organisations that move quickly to adjust the approach to search marketing, measurement, and content will see results faster than those still relying on traditional SEO playbooks”.
Google has thrown another grenade into Search, extending the AI features that are already making life harder for publishers, website owners, and marketers. Announced as part of its annual I/O developer conference, the new changes revolve around transforming the classic Search box into an “intelligent” window, one which expands according to the user’s queries and which can process multi-step prompts. Google has also integrated its now-familiar AI Overviews feature with AI Mode, which users can seamlessly use to input further queries after receiving an AI-generated summary.
The Big Tech giant is calling this “the biggest upgrade to our Search box in over 25 years”. In addition, it’s rolling out Search agents, which ”operate in the background 24/7” and provide regular updates on and summaries of topics the user has asked to track. The idea here is that the user will have to perform fewer manual searches themselves, and while it does look like a potential timesaver, it’s already beginning to unnerve SEO teams the world over.
Traditional optimisation models are rapidly becoming redundant
