For two decades, the arrangement between Google and the open web was straightforward enough that nobody bothered to name it. Publishers produced the content, Google indexed it, and users clicked through to read the rest. But today, that arrangement has ended, and the data from Chartbeat, Seer Interactive, and the Reuters Institute make it difficult to argue otherwise.

The strange part is how many B2B teams are still operating as though the rules from 2022 continue to apply. Rankings get measured, position-one wins get celebrated in weekly reports, and AI Overviews get treated as a temporary nuisance instead of what the numbers show them to be. 

If you’re facing something similar, you might want to read this one.

In March 2026, Chartbeat released the dataset that finally put hard numbers on what publishers had been whispering about for 18 months. 

Over a two-year window, referral traffic from traditional search engines declined 60% for small publishers, 47% for medium publishers, and 22% for large publishers. 

Page views from Google Search fell 34% between December 2024 and December 2025, and Google Discover dropped 15% in the same period.

ChatGPT referrals grew more than 200% in the same window, which sounds dramatic until you see the baseline. Chatbots still account for less than 1% of all publisher page view referrals, which means the much-discussed AI traffic replacement is, for now, a rounding error on a ledger that is otherwise bleeding out.

What we are seeing today is a redistribution of attention away from the open web and into AI answer interfaces that were never built to send anyone anywhere in the first place. The problem isn’t that AI replaced Google, but that Google replaced itself.

Traffic declines are only the first layer of the problem. The deeper issue, and the one most content teams have not reported honestly to their executives yet, is that you can rank in position one and still lose most of your clicks the second an AI Overview sits above your result.

Seer Interactive ran the study that I think will eventually be cited as the definitive one, analyzing 3,119 informational queries across 42 organizations between June 2024 and September 2025.

When an AI Overview appeared, organic CTR collapsed from 1.76% to 0.61%, a 61% drop on queries where ranking positions held steady. Paid CTR fell further, from 19.7% to 6.34%, which shows a massive 68% decline.

The Mail Online shared similar internal numbers with Press Gazette, reporting that organic CTR averaged 13% on desktop and 20% on mobile when no AI Overview was present, and dropped to 5% and 7% respectively when one appeared at the top of the page. 

The UK Professional Publishers Association went further when it submitted evidence to the Competition and Markets Authority, citing a lifestyle publisher whose CTR on a popular query fell from 5.1% to 0.6% over the past year, despite its page continuing to rank throughout.

The data clearly shows that rankings no longer map to traffic. And this sentence should probably be printed on the wall of every content team still reporting to executives in terms of keyword positions and SERP movement.

There is, buried in the same Seer dataset, something that looks a lot like a survival path. Brands cited inside AI Overviews earn 35% more organic clicks and 91% more paid clicks than brands that are not. This means that the gap between being in the answer and being beneath it has widened dramatically. 

Research from Princeton, presented at ACM SIGKDD in 2024, found that content optimized specifically for generative engines can lift AI visibility by up to 40%.

This is where the industry has started to converge on a new discipline, call it Generative Engine Optimization, Answer Engine Optimization, or whatever acronym outlasts the others. The shift underneath all the names is the same one. You are no longer competing for a position in a list of blue links. You are competing to be one of the three or four sources an AI model pulls into its synthesized answer, and everything else you are doing with content is increasingly downstream of whether you make it into that answer.

Gartner has predicted that traditional search volume will drop 25% by 2026. Walker Sands reports that 90% of B2B buyers now integrate generative AI somewhere in their buying journey. 

The 2026 Reuters Institute survey, drawn from 280 media leaders across 51 countries, flagged AEO and GEO as two of the terms it expects to become mainstream this year. If your content is not structured for citation, what you are losing is not just traffic. You are losing presence in the rooms where vendor shortlists get built before anyone has visited your website.

The Princeton research highlighted something uncomfortable for anyone whose SEO program is still built around keyword density and link acquisition. Those tactics performed poorly in generative environments, and what worked instead was different in kind rather than degree. Embedded citations lifted AI visibility considerably, as did statistics with named sources, authoritative language, and clear entity signals spread consistently across a site.

Traditional SEO was designed for crawlers matching keywords to queries. AI systems are designed for extraction, which is a different problem entirely. They are looking for direct, atomic, verifiable facts that can be pulled into a synthesized answer and attributed cleanly to a source. 

A page that spends 500 words on context before arriving at the point will still rank fine on Google, and will be ignored almost entirely by ChatGPT. A page that leads with the answer in the first 40 to 60 words, followed by supporting evidence, gets cited at much higher rates across every AI engine that has been studied so far.

The format of your content has become a ranking factor in a system where ranking and traffic are no longer tied to each other in any reliable way. This is why a lot of the SEO work being done in 2026 is quietly shifting underneath the surface. 

Content teams are restructuring top-performing pages to lead with direct answers, adding citation-grade statistics with named sources, and building topic authority through consistent entity signals across a domain rather than chasing individual keyword wins one post at a time.

The organizations that are going to come out of this period in a strong position are not the ones that panicked and slashed their content budgets. They are the ones who reframed what content was actually for in the first place.

If traffic was the goal, the game is harder than it has ever been, and probably harder than it will ever be again. If influencing inside AI answers is the goal, the field is more open than it has been in a decade, because most brands have not yet adapted. 

Foundation's internal data suggests that even 5% citation share from your own domain qualifies as a significant win in competitive B2B verticals, with the rest of the influence coming from being stacked across Reddit, YouTube, review platforms, and third-party publications that AI engines pull from consistently.

What this changes practically is the process of keyword research itself. Keywords still matter, but they sit alongside prompt research now. What are buyers actually asking ChatGPT and Perplexity about your category, which competitors are getting cited in the answers, and which topics are still yours to claim because nobody has structured authoritative content around them yet?

The B2B buyer journey in 2026 often begins well before anyone visits a website, and a VP of Marketing asking an AI assistant for the best demand generation platforms is never going to see your landing page if your brand is not already inside the synthesized answer. 

The content investment that gets you into that answer looks different from the content investment that used to get you to the top of a search results page, and the teams figuring that out now are buying themselves a two-year head start on everyone who is still waiting for the dust to settle.

SEO isn't dead, and the framing that it is tends to be lazy and mostly sold by people trying to rebrand the same service under a new acronym. What has actually died is a specific bargain between search engines and the web, where good content earned traffic in exchange for being indexed, and that bargain had a twenty-year run before it quietly expired.

The new bargain is harder and considerably less generous. Good content now earns citation, and citation is worth something, but it is worth less than traffic used to be worth for most businesses that built their models on organic acquisition. 

Companies that understand this are already rebuilding their content operations around authority, extractability, and citation tracking. Companies failing to do so are watching their organic traffic numbers decline quarter over quarter and filing it under market conditions or seasonality.

The 2026 data makes one thing clear enough that it should no longer be controversial. Rankings are no longer a reliable proxy for anything that matters to a business. And it’s about time content teams start measuring what actually influences buying decisions in 2026.