.png){kind=logo}
For twenty years, companies paid billions to rank on Google. A growing number of people are now getting their answers from AI assistants that do not return links at all. Nobody has fully figured out what that means for the businesses that built empires on search traffic.
The discipline of search engine optimisation was built on a specific model of how information flows across the internet. Search engines crawl documents, score them for relevance and authority using hundreds of ranking signals, and return a list of links ordered by estimated quality. The entire industry that grew around this model, worth tens of billions of dollars annually, exists to influence where a given document appears in that ordered list.
The model worked because it accurately described how people found things online. Type a query into Google, receive a list of links, click the most relevant one. For the better part of two decades, that sequence described the majority of information discovery on the internet.
That sequence is now changing.
AI assistants have introduced a different interface for accessing information. When a user types a question into Perplexity, ChatGPT or Google Gemini, the system does not return a ranked list of documents. It constructs a response. The answer typically runs to several paragraphs, may cite two or three sources, and represents the system's best synthesis of what the relevant information landscape broadly agrees on.
Mohit Ahuja, founder of distribution infrastructure startup Ampli5, refers to this dynamic as the “Answer Consensus Mechanism.”
To understand why the old playbook breaks down, it helps to understand what AI assistants are actually doing when they generate a response.
Large language models are trained on vast datasets of internet text. In the process, they develop implicit representations of which information is reliable, which positions are widely held and which claims appear consistently across unrelated sources. This is not the same as a search ranking. It is closer to an approximation of collective knowledge, weighted by how often a given position appears across the distributed information ecosystem.
When an AI assistant with retrieval capability, such as Perplexity or a ChatGPT instance with browsing enabled, responds to a query in real time, it performs a version of the same operation dynamically. It reads across multiple sources and synthesises a position that reflects what the current information environment broadly supports.
Research conducted by Ampli5 across dozens of product categories found that AI assistants draw heavily from sources that have historically been low priorities for SEO teams: Reddit communities, YouTube video transcripts, podcast content, expert blogs, niche forums and community discussion threads. Brand websites and corporate content, the primary focus of most SEO investment, were cited comparatively rarely.
The pattern reflects the underlying logic of how these systems assess reliability. A claim made on a brand's own website is a single source with an obvious interest in the outcome. The same claim appearing independently across a forum discussion, a creator video, an expert's published commentary and a media article begins to resemble something more like consensus. For a system trained to approximate collective knowledge, that distinction is significant.
Mohit Ahuja calls this dynamic the Answer Consensus Mechanism. The term describes the tendency of AI systems to preferentially surface information that appears consistently across multiple independent, contextually relevant sources, rather than information concentrated at a single authoritative URL.
"These systems are not looking for the best page," he says. "They are looking for the most agreed-upon answer. Those are different things, and they live in different places on the internet."
The broader challenge for the industry is that nobody has agreed on how to measure AI visibility, let alone how to influence it.
Traditional SEO has mature measurement infrastructure. Rankings can be tracked. Traffic can be attributed. The relationship between ranking position and organic traffic is well-documented. A company that drops from position one to position three on a competitive keyword can calculate the expected traffic impact to within a reasonable margin.
AI visibility has none of that infrastructure yet. A brand that is being cited inside ChatGPT responses has limited visibility into when, in response to which queries, or how often it is mentioned. A brand that is absent from AI-generated answers has almost no way to diagnose why, or to measure the commercial cost of that absence.
This creates an uncomfortable situation for marketing and growth teams at consumer and B2B companies alike. The evidence that AI assistants are absorbing meaningful volumes of discovery intent is accumulating rapidly. The tools to respond to that shift are still early.
Perplexity has reported hundreds of millions of queries per month. OpenAI reported over 300 million weekly active users for ChatGPT earlier this year. Multiple independent analyses of search behaviour have identified a measurable decline in traditional search queries among users under thirty-five, with AI assistants absorbing a growing share of the discovery phase of purchase decisions.
For categories where decision-making matters, including technology, health, financial products, travel and professional services, the question of what AI assistants say about your brand is no longer speculative. It is a distribution problem with real commercial consequences.
Whether or not Ampli5's specific approach proves durable, the underlying problem it is addressing is real and growing.
The marketing and growth industry is reaching the end of a long period in which the rules of online visibility were relatively stable and well-understood. A new period is beginning, with different rules that are not yet fully understood by anyone, including the companies building the AI assistants that are driving the shift.
Google itself is navigating this tension. Its AI Overviews product, which generates synthesised answers at the top of search results pages, has drawn significant attention from publishers and brands concerned about its implications for organic traffic. The feature changes the relationship between a search query and a website visit in ways the SEO industry is still working to understand.
Industry observers increasingly view this shift as the emergence of a new infrastructure layer for the internet.
Search engines organized information through ranking.
AI assistants organize it through synthesis.
If that trend continues, the companies that shape how information appears across distributed platforms may wield more influence over discovery than those that simply optimize websites.
You can get your own free demo of what questions pertaining your business is not answered on the internet at Ampli5
