For years, digital visibility followed a simple rule. Rank high on Google, and you were discoverable. Be discoverable, and you had a shot at winning the deal. It was a system that, while competitive, was ultimately predictable.
That equation no longer holds.
As AI-powered search reshapes how buyers research and evaluate companies, a growing disconnect is emerging between where brands rank and whether they’re actually considered. And in many cases, the gap is bigger than most teams realize.
The shift isn’t gradual. According to Shane H. Tepper, cofounder of Resonate Labs, a company specializing in AI search, it becomes obvious the moment companies stop relying on rankings as a proxy for visibility and start looking at how AI systems actually behave. “It breaks the moment you test it against AI citation data instead of taking it on faith,” he explains.
Data backs that up. Analysis across more than 50,000 AI prompts shows that traditional SEO signals barely correlate with whether a brand gets cited in AI-generated answers. Organic traffic explains about 5% of citation behavior. Backlinks account for even less. In practical terms, the metrics that have guided search strategy for years have almost no predictive power in AI environments.
That’s because the underlying system has changed. Search used to operate on a single, visible scoreboard where rankings directly influenced clicks. Today, buyers interact with multiple AI surfaces, each with its own logic for retrieving, synthesizing, and prioritizing information. The rules no longer translate.
A recent study from Semrush found that 90% of sources cited in AI-generated answers don’t come from top-ranking search results. But the deeper shift isn’t just that rankings matter less. It’s that entirely different signals are taking their place.
Only around 40% of AI citations come from Google’s top ten results, while a meaningful portion of cited pages have no traditional search visibility at all. AI systems are surfacing content that, by SEO standards, would be considered invisible.
So what determines who gets cited?
Tepper points to a consistent pattern across multiple analyses: AI systems prioritize content that is specific, structured, and easily extractable, while also favoring brands that are well represented across the broader web. “The pattern across multiple independent analyses is consistent: what matters is whether your brand is well-represented across the web in the kinds of content AI models trust,” he says.
That includes clear claims, structured information, and data points that can be pulled directly into an answer. In this environment, a page with little traffic and few backlinks can still be cited, while a top-ranking page can be ignored entirely.
At the same time, the way buyers discover companies is shifting upstream. Tools like ChatGPT are increasingly used to compare vendors, evaluate trade-offs, and form early opinions, often before a traditional search ever happens. While Google’s AI Overviews are already reducing click-through rates, the more significant change is happening before search even begins.
Buyers are now spending time inside AI interfaces conducting research that leaves no trace in analytics. A typical journey looks different: a buyer evaluates options within an AI tool, narrows down choices, and only later searches for a specific brand. What shows up in analytics is a branded search or direct visit. What doesn’t show up is the process that shaped the decision.
Tepper refers to this as “dark traffic,” a form of influence that exists but leaves no measurable signal. “A buyer spends 25 minutes in a ChatGPT conversation comparing five vendors… Your analytics records a branded search. The 25-minute session that generated the intent is gone.”
That invisibility creates a new kind of risk. For teams still heavily invested in traditional SEO, the issue isn’t an immediate drop in performance, but something harder to detect. “You maintain your rankings, your organic traffic looks stable, and meanwhile buyers are building shortlists inside AI conversations where you don’t appear,” Tepper says.
In competitive markets, that early-stage visibility matters more than ever. If AI systems are helping shape shortlists before companies enter the sales process, then not appearing in those conversations means losing opportunities before they even begin.
The challenge is compounded by fragmentation. Different AI platforms surface different sources, with limited overlap between them. Visibility in one system doesn’t guarantee visibility in another. Companies that continue optimizing solely for Google are effectively competing on a single surface while buyers are making decisions across several.
At the same time, the measurement gap makes the impact difficult to see. AI-influenced research often shows up as branded or direct traffic, giving the impression that existing strategies are working even when the underlying drivers have shifted.
None of this makes SEO irrelevant. But it does make it insufficient on its own.
The focus is moving from ranking to representation — how a company is described, cited, and recommended across AI-generated responses. That requires a different approach to content, one built around clarity, specificity, and presence across trusted sources.
Because in an environment where buyers can research, compare, and decide without leaving an AI interface, visibility is no longer defined by where you rank.
It’s defined by whether you’re included in the conversation at all.