Signal Through the Slop — Week 35, August 2026

AI search is changing how brands get found, recommended, and trusted. Each week, we pull signal from Reddit, LinkedIn, and industry research to track what's actually moving in the space. Here's what caught our attention this week.

What Reddit is saying this week

The loudest conversation on Reddit this week isn't about what to do in AI search — it's about what just happened to a platform AI search depends on. ChatGPT Search dropped its citation rate for Reddit content by roughly 94% over six days following August 14, going from about 250 Reddit citations per day to just 16. Over the same period, Perplexity roughly doubled its Reddit citations. Three major platforms, three completely different behaviors — and practitioners trying to build strategies for "AI search" are realizing they've been optimizing for a category that doesn't actually behave as one.

The more persistent thread this week is the traffic attribution problem. Practitioners everywhere are watching organic traffic decline and struggling to explain it to leadership cleanly. Is it the AI Overviews expansion? The August spam update? An algorithm shift unrelated to AI? The frustration isn't the decline itself — it's that the standard reporting tools don't make it easy to separate "AI Overviews took our clicks" from "we made a content mistake." This gap between what's happening and what teams can prove is showing up in thread after thread with no clean answer.

The rank-versus-citation disconnect is getting surfaced in multiple subreddits this week. Practitioners who have strong Google rankings are reporting that they're invisible in Perplexity answers and sparse in ChatGPT's citations for the same queries. New research from Machine Relations this month confirms what they're experiencing: 80% of URLs that AI platforms cite don't rank in Google's top 100 for the same query. The two signals are measuring different things. SEO rank and AI citation aren't the same problem, and most teams don't have a separate program to address the difference.

The organizational ownership question is also running hot. Who owns generative engine optimization — the SEO team, content, PR, or brand? Reddit doesn't have a consensus answer, and neither does anyone else. The threads are full of "it depends on your org" responses, which is functionally useless for someone who needs to write a job description or get budget approved. The question of who owns it won't resolve until the measurement system is defined — and most teams haven't done that yet.

What LinkedIn is saying this week

The Conductor State of AEO/GEO CMO Investment Report is making the rounds this week, generating real engagement from senior marketing buyers. The headline number — 94% of CMOs plan to increase AEO/GEO investment in 2026 — is getting shared widely. But the more interesting signal is in the comments, where CMOs are asking questions like "how do we know if our content is actually getting cited versus just appearing in AI answers?" and "how do we structure the team to own this?" The investment decision has been made; the operating model hasn't caught up.

Google's VP of Search told CMOs at a recent event that "Good SEO is good GEO" — and it generated substantial pushback on LinkedIn from practitioners who know the two require different signals, different content formats, and different measurement approaches. The 80% non-top-100 citation finding is the clearest rebuttal: if good SEO automatically translated to AI visibility, the data wouldn't show what it shows. The platforms surface different content for different reasons, and assuming parity is now a documented mistake.

Search Engine Land published a piece this week titled "AI search is here, but marketers still call it SEO" that's surfacing useful data about how buyers think about the problem. Seventy percent of marketers would search for "AI search optimization" or just "SEO" when looking for help with AI search visibility — even as CMOs use "GEO" at three times the rate of individual contributors. The vocabulary gap between buyers and practitioners is real, and it's creating friction in every direction: agencies pitching services, buyers evaluating them, and teams trying to build consensus internally.

Google's announcement of a Search Console toggle to let publishers opt out of AI Overviews and AI Mode without ranking penalty is generating genuine strategic conversation on LinkedIn. The opt-out framing is new — and it's creating a decision moment for brands who haven't thought carefully about whether being cited in AI Overviews is beneficial or harmful for their specific business. Most brands don't know the answer yet, because most brands don't know what AI Overviews are currently saying about them.

What the research shows this week

The headline data point this week is from Machine Relations' August 2026 research: 80% of URLs cited across ChatGPT, Perplexity, Copilot, and Google AI Mode don't rank in the top 100 Google results for the same query. This dismantles the assumption that strong SEO translates automatically to AI visibility. The two signals are measuring different things — entity authority, third-party citation presence, structured content format, and platform-specific retrieval behavior all drive AI citation in ways that traditional rank factors don't capture.

Semrush's expanded AI Visibility Index — built from 126 million U.S. AI search prompts — surfaced a distinction that most brands are getting wrong: being mentioned in an AI-generated answer is not the same as having your site cited as a supporting source. A brand can appear in 40% of AI Overviews covering a category while having its website cited in fewer than 5% of them. These are different metrics, and measuring the wrong one produces the wrong strategy.

Google moved fast on product infrastructure this week. Gemini 3.7 Flash became a selectable model in AI Mode on August 14, bringing improved instruction following and intent understanding to AI Mode answers for Pro and Ultra subscribers. The August 2026 spam update followed on August 18-21 — Google's third spam update this year — and it's the first to explicitly cover attempts to manipulate generative AI responses in AI Overviews and AI Mode, a policy Google clarified in May 2026. Google also confirmed it's testing a Search Console toggle for AI Overviews and AI Mode opt-out, which won't affect standard search rankings.

The zero-click picture keeps sharpening. Zero-click searches have reached 64.82% overall in 2026; when AI Overviews appear, organic CTR drops roughly 61%. But clicks that survive are converting 23% better than pre-AI Overview baselines — users who click through after reading an AI summary already understand the content and are seeking depth or action. The volume-versus-quality tradeoff is real, and it's increasingly the frame that makes AI search investment legible to leadership.

That's the signal this week — back next Monday with more.