Google’s AI Mode passed one billion monthly users within a year of launch, and query volume has more than doubled every quarter since, according to figures the company shared at its 2026 developer conference. For a search engine that has organised the web around ranked blue links for over two decades, that is a structural shift in how people find information, not a feature update.
For most Australian businesses, the consequence arrives quietly. A share of the customers who used to click through to a website now get their answer inside the search results page, or inside a conversational exchange with an AI system, and never visit at all. The website still exists. The keyword may still rank. The visit simply stops happening.
This is not a marginal effect confined to media publishers, though publishers were the first to notice it and the loudest to complain. It reaches any business whose growth plan depends on organic search: professional services firms whose comparison and “how to choose” content used to convert cold visitors, retailers whose category pages drove consideration, and every marketing function still treating a first-page Google ranking as the finish line rather than one input into a larger visibility problem.
The Numbers Behind the Shift
The scale of the change is now documented, not anecdotal. Pew Research Center tracked the on-platform behaviour of 900 US search users across roughly 69,000 real Google searches in early 2025, and found that when an AI summary appeared above the results, users clicked through to a website in 8 percent of visits, compared with 15 percent when no summary appeared. Sessions with an AI summary present were also more likely to end there entirely: 26 percent of those visits saw no further search activity, against 16 percent without one. Pew’s full analysis is one of the first studies to measure this from real browsing panels rather than survey responses.
Ahrefs reached a comparable conclusion from a different angle, analysing 300,000 keywords before and after the US rollout of AI Overviews and finding a 34.5 percent average drop in click-through rate to the top-ranking result once an AI Overview appeared for that query. Positions that once reliably converted attention into a visit are quietly converting attention into a paragraph the user never has to leave Google to read. The full dataset is published here.
Australian businesses are not exempt from either trend. Google’s AI Mode launched locally in October 2025 and now sits alongside AI Overviews in standard search results, and Australian publishers have already reported the same quoted-without-visit pattern affecting their own traffic. The competitive pressure compounds the effect: OpenAI and Perplexity are both explicitly positioning their own products as substitutes for a Google search box, drawing a share of high-intent commercial queries away from search entirely, before any ranking algorithm gets involved.
From Ten Blue Links to One Answer
The mechanical change matters more than the branding. AI Overviews and AI Mode do not simply display a search result higher on the page. They synthesise a response from several sources, present it as a single answer, and increasingly invite a follow-up question that keeps the entire exchange inside Google’s interface. At its 2026 developer conference the company described this as its biggest search box redesign in over twenty-five years, moving to Gemini 3.5 Flash as the default underlying model and enabling a continuous conversational thread rather than a series of disconnected queries.
The practical effect for a business is that ranking first is no longer sufficient on its own. What matters increasingly is whether a business’s facts, claims, and positioning get selected as source material for the answer itself. A page can rank on page one and still contribute zero visits, if the AI system extracts the answer and never surfaces the link with enough prominence to earn a click. Visibility has split into two separate outcomes: appearing in the answer, and being visited because of it. Most marketing programs are still built to optimise only the second one.
What Actually Changes for a Marketing Function
The instinct to declare SEO dead and replace it with something new is premature and, for most businesses, a distraction. The technical foundations that made a website legible to Google, clean information architecture, fast load times, genuine topical authority, still make it legible to an AI system, because these systems still crawl and index the same web. A technical SEO audit remains the correct starting point, not a legacy exercise.
What changes is the emphasis placed on top of that foundation. Structured data, in the form of schema.org markup, stops being a minor ranking signal and becomes a primary way a language model confirms facts about a business with confidence: its services, its credentials, its pricing structure, its location, its reviews. Content that states a claim once, clearly, and attributes it to a source is easier for a model to extract and cite than content that builds toward a conclusion across six paragraphs of narrative framing. Original data, named case studies, and clearly dated facts function as citable primary material in a way that generic, thinly rewritten advice does not.
The broader digital footprint also carries more weight than it used to. Independent citations across directories, review platforms, industry publications, and press coverage act as corroborating signals that a language model weighs when deciding whether to trust and repeat a claim about a business. This is closer to the logic of a Wikipedia entry than a backlink profile: consistency and independent corroboration across many sources matter more than the authority of any single link.
The Budget and Attribution Problem
For a chief executive or a chief financial officer, the more uncomfortable implication sits in measurement, not tactics. Last-click and last-touch attribution models already understated the contribution of brand and content marketing before generative search existed. They understate it further now, because a meaningful share of influence happens inside a platform interface the business does not control and cannot instrument with an analytics tag. A prospective customer can read a synthesised answer that favourably mentions a business, form a preference on the strength of it, and only convert weeks later through a branded search or a direct visit that a dashboard attributes to an entirely different channel.
Brand and product mentions inside AI-generated answers are becoming a leading indicator that most marketing reporting does not yet capture, in the same way that share of voice mattered in traditional media long before digital attribution existed. Boards and finance functions evaluating marketing spend against click-based metrics alone are increasingly measuring a shrinking part of a larger picture, and risk cutting the programs, content, and organic search investment that are actually building the visibility a click-based report can no longer see.
A Practical Framework for the Next Two Quarters
None of this requires abandoning an existing marketing program. It requires widening what that program measures and produces. Five priorities are worth acting on before the next budget cycle:
- Audit current visibility inside AI answers, not just search rankings. Test the specific questions your buyers actually ask across Google’s AI Mode, ChatGPT, and Perplexity, and record whether your business is mentioned, misrepresented, or absent entirely.
- Treat structured data as infrastructure, not decoration. Complete and accurate schema markup is how a language model confirms facts about a business rather than guessing at them from unstructured text.
- Write to be extracted, not just read. Direct answers, clearly defined terms, and stated data points are more likely to be lifted into an AI-generated response than narrative copy that delays its conclusion.
- Add brand mention tracking to the marketing dashboard. Share of voice inside AI answers deserves a place alongside rankings, traffic, and conversion rate, informed by the same audience and market research that already shapes positioning.
- Treat digital PR and independent citation building as core infrastructure. Reviews, directory listings, and third-party coverage now function as corroborating signal for AI systems in addition to their existing reputational value.
Where This Leaves Business Leaders
The shift toward AI-mediated search is not a hypothetical to plan for later. It is already measurable in click-through data, already live in the Australian market, and already reshaping the competitive landscape as OpenAI and Perplexity court the same commercial queries Google has held for two decades. Businesses that keep measuring visibility only through rankings and click volume will keep making decisions on an incomplete picture, precisely as that picture becomes less complete every quarter.
The businesses that adapt fastest will not be the ones that abandon search discipline. They will be the ones that extend it: pairing the technical rigour that has always underpinned good SEO with a deliberate strategy for how a business gets represented, cited, and recommended inside a growing share of searches that never produce a visible list of links at all. A digital marketing program built for that reality looks different from one still optimised purely for the ten blue links, and the businesses that update it first will hold that advantage for longer than the technology stays new.