Google just made the biggest change to search in 25 years, and most of marketing is still working as if it didn’t. At its I/O summit, the company confirmed what the data already showed: search is moving from a list of blue links to a conversational AI answer box. Those blue links have been the front door to the internet for a generation of marketers, and that door is closing. A lot of teams are still doing SEO like it’s 2015.
The numbers are already ugly. Zero-click searches now make up 68% of all Google queries, and Google desktop searches per user have fallen nearly 20% year over year across tens of millions of U.S. users. The channel most brands built their reach on is already breaking. Referral traffic is drying up, and the conversations that decide how people see a brand are moving to social, the place most teams watch least.
So the definition of a marketing win has to change too. It used to be sessions and click-through rate. Now it’s whether your brand shows up, gets trusted, and gets described accurately wherever answers get assembled, even when nobody ever lands on your site. That’s a different job than most stacks were built for. Pretend otherwise and good brands quietly get written out of the story.
The channel marketers trusted is draining
It’s already in the traffic logs. The New York Times watched organic search fall to 36.5% of its desktop and mobile traffic in April 2025, down from 44% three years earlier, per Similarweb figures reported by the Wall Street Journal. If one of the strongest publishers on the web is losing that much ground, smaller brands should assume the same thing is happening to them.
Google knows it too. In its antitrust fight with the Justice Department, the company conceded in a court filing that the open web is already in rapid decline, a very different story than the one its executives tell in public. When the biggest gatekeeper of web traffic tells a judge the open web is fading, that’s worth believing.
The industry’s answer so far is a new set of acronyms. Generative engine optimization and answer engine optimization, GEO and AEO, both about staying visible inside AI answers. They matter, but they work on the last step in the chain and ignore where the machine gets its material in the first place. Interestingly, social media is now a major source for AI platforms, optimizing for AI responses without watching social is polishing the output while ignoring the source.
Search moved into the feed
People are still searching for things online, they hey just started looking beyond Google to social. Nearly half of U.S. consumers now use TikTok as a search engine , up from 41% the year before, and among Gen Z the amount jumps to 64%. When someone wants a restaurant, a product, or a read on a brand, they type it into a feed before they ever open Google.
Social media feeds both ends of discovery. It’s where people search, and it’s the raw material AI pulls from behind the scenes to deliver query answers. Deloitte found that evaluative research and comparisons are increasingly AI-mediated, with generative systems leaning hard on social content and reviews to build their recommendations. So the story people tell about your brand on social becomes the source material for the answer a machine gives someone who’s never visited a single one of your channels. Your next customer might form a full opinion of you from a video on TikTok you never saw, summarized by AI answer you’ll never read.
The first move for any marketing leader is to accept that this is already underway. The second is to find a way to see where their brand narrative is forming, which is harder because it means reevaluating the set of tools and methodologies needed for monitoring and bolstering brand perception.
Marketing Technology News: MarTech Interview with Mark Listes, CEO @ Pendulum Intelligence
Video is the part marketers must read
The dominant format online is also the one almost no marketing tool can analyze: video. Pew’s 2025 Social Media Fact Sheet shows that YouTube reaches 84% of U.S. adults and TikTok 37%, with both audiences still growing. Meaning now lives in spoken audio, on-screen text, tone, and visual context. One person with a camera can change how millions feel about a brand without ever typing its name.
Legacy social listening tools can’t see any of it, because they was built for text. They read written posts, pull keywords, score sentiment against strings of words. This works fine on a tweet, but becomes blind the second the conversation becomes a 15-second clip where the tone carries the message and the caption is blank. “This product is great!” means praise in a real recommendation and the opposite in a fail compilation, and a keyword tool scores them the same. Most brands end up paying for dashboards that are confidently wrong about brands’ perception.
Reading video takes three things most stacks don’t have. First, speech-to-text at real platform scale, across slang, languages, and messy informal speech, not tidy caption feeds. Second, multimodal sentiment, so tone and visual cues get read next to the transcript and sarcasm doesn’t land as applause. And lastly, narrative-level pattern detection, so you can tell when the same claim is spreading across independent creators fast, not just score one video at a time. Put together, they let you watch a brand story form where it actually forms, on screen.
Audience reaction is the new proof
There’s a bigger shift underneath all this, and it’s about proof. In a feed full of synthetic content, you can’t prove something is real or know which authentic narratives are shaping around your brand by looking at the content anymore, because anyone can fake a convincing video now. What you can still see is whether real people reacted to it, and how. Genuine audience response, built up over time, is the one signal nobody can fake at scale. Right now it’s the closest thing to proof we’ve got.
That changes how a leadership team should think about trust. The job is to watch how real audiences behave around your content, over time. Chasing individual posts for “AI-generated” labels is a losing game, instead trust shows up in a track record. When your marketing says one thing and your product does another, people catch the gap fast, and now they catch it on camera.
The marketers who win the next few years are the ones who can read the room, wherever the room happens to be. That’s the line between the brands that get to write their own story and the ones who let everyone else write it for them.
About The Author Of This Article
Chen Guter is CMO of dig
About dig
dig is a leader in video-first social intelligence, helping brands, agencies, and governments understand the authentic human reactions influencing perception across social.
Marketing Technology News: How MarTech Is Enabling Autonomous Brand Engagement Across Channels?
