The end of browsing: How ChatGPT is collapsing the marketing funnel
Chris Andrew • Imagine you’re looking for a chicken soup recipe using Google search: You click on the first result, and instead of a recipe, you get the backstory of a grandmother's secret recipe. You scroll past ad placements, autoplay videos, 2,000 words of meandering narrative, and 62 instances of the word “chicken” before you finally arrive at the recipe.
This was the internet of browsing, and things were getting dire. It relied on a quid pro quo: Publishers gave away content for free, packed it with ads and keywords, and in return captured traffic they could monetize. Users paid with their time, their attention, and their data: scrolling past distracting pop-ups and cluttered pages to find what they were looking for.
The longer this system went on, the less well it worked. Scrunch tells the story of how AI upended the old way of browsing and what brands can do to adapt.
For brands, the golden era of cheap traffic arbitrage ended years ago. Paid ads that once delivered outsized returns in the mid-2010s had become more expensive and less reliable. Rising cost per clicks on Google, climbing cost-per-thousand impressions on Facebook, and privacy changes that made attribution harder all meant marketers were paying more to reach the same audiences. To compensate, publishers and brands doubled down on squeezing value out of every visitor.
For users, that translated into an internet experience that got steadily worse. Pages were stuffed with more ads, keyword-stuffed copy, annoying exit pop-ups, and manipulative design patterns—all in the service of keeping people on-site a little longer or nudging them toward a conversion (not to mention helping Google collect AdSense revenue along the way). It was a race to the bottom where browsing felt increasingly like wading through sludge just to get to the content you wanted.
Then along came large language models (LLMs). Whether a user was engaging directly with ChatGPT or typing a query into Google and getting an AI Overview (AIO) in response, the result was the same: a clear, concise answer in seconds. No ads, no keyword padding, no bait-and-switch … and no more quid pro quo. Once people taste that simplicity, it’s hard to go back to clicking through crowded web pages. The browsing era of the internet is ending.
For users, this is mostly a win: They no longer have to endure endless selling and distraction to get the information they need. For publishers and brands, it upends the model they’ve relied on for decades. But it also provides an opportunity to reach fewer visitors who are much closer to buying—more on that below.
We’re already seeing the effects: Across the board, big publishers have seen steep declines in organic traffic. According to SimilarWeb, overall organic traffic to news publishers dropped by over 25% between May 2024 and May 2025. Major outlets such as CNN, Business Insider, and HuffPost reported decreases ranging from 28% up to 40% in the same period. Even HubSpot—the long-time gold standard for B2B SEO—has seen its blog traffic collapse by around 70%-75% since Google rolled out AI Overviews. While some analysts attribute this to search algorithm updates, the more straightforward explanation is that AI-powered answers have started siphoning away the browsing traffic that once sustained publishers and brands. And when that traffic disappears, it’s not just pageviews that vanish—it undermines the entire funnel marketers once depended on.
The old marketing funnel depended on organic discovery: people searched, clicked through to your site, got retargeted, filled out a form somewhere along the way, and eventually converted. But when ChatGPT, AIO, and other AI platforms deliver answers directly—and users aren’t required to click for answers—that funnel collapses.
To survive in this new environment, marketers need to start thinking less in terms of user traffic and more in terms of AI comprehension. Can ChatGPT clearly articulate what your company does, who it serves, and why it’s better than alternatives? Does Perplexity cite your product page, or a competitor’s? Is Gemini referencing an outdated blog entry? The brands that endure will be the ones that treat AI platforms as their most important audience. That means monitoring how they appear across platforms, fixing technical blockers that prevent AI from retrieving their content, and deliberately publishing information in formats AI can parse. In other words: optimizing not for clicks from users, but for AI model comprehension.
The brands that crack this code might see less traffic overall, but the traffic they get will be far more valuable (i.e., higher quality and higher converting). Increasingly, users arrive at brand sites not to browse but to buy. They’ve already outsourced the research phase to the foundational large language models, and they show up ready to complete a transaction. In fact, data from Scrunch AI shows that visitors referred from AI search platforms are three to five times more likely to convert than those from traditional organic search, reflecting the much higher intent of these visitors. Brands should be optimizing for this kind of traffic—not distracting purchase-ready buyers with ads, popups, lead magnets, and fluffy copy.
The end of browsing isn’t the end of discovery. It makes discovery easier and more efficient for consumers, who never volunteered to scroll past pop-ups, autoplay videos, and keyword-stuffed paragraphs to get an answer. Without the incentive to clutter pages with ads and distractions, what’s left is a cleaner, more direct internet: users get better experiences, and brands meet them at the moment of highest intent. The brands that stop mourning lost clicks and start optimizing for AI comprehension will be in a stronger position than before: fewer visits, but far better ones.