Signal Through the Slop — Week 41, October 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

Thread-level signal was thin this week, so the more useful read is on Reddit's role in AI answers rather than any single discussion. New data keeps putting the platform at the center of how answer engines source information. AthenaHQ's report, published September 29, found Reddit accounts for 21.9% of external citations across eight LLMs, more than double the next source. Search Engine Journal's September 30 roundup adds that Ahrefs counts 5.8 million Reddit citations from Google AI Overviews alone, and SISTRIX has reddit.com as the fourth most visible domain in U.S. Google Search.

That makes the question practitioners keep asking pretty predictable: if Reddit is this influential, what is a brand supposed to do about it? The honest answer from the data is that earned, genuine participation is the lever, not promotion. Reddit reaches 59% of U.S. business decision-makers according to Comscore and Reddit research cited in the SEJ piece, so the audience and the citation value overlap in a way few channels do.

If you're managing AI search visibility for a brand, the thing worth doing this week is an audit of where your category's questions are already being answered on Reddit, and whether anyone credible is showing up in those threads. The topic worth covering: how to participate in communities in a way that earns citations without getting burned.

What LinkedIn is saying this week

Measurement dominated the professional conversation. The sharpest post came from Nick Chasinov, who argued that the one-dashboard dream for AI search doesn't exist, and laid out six separate signals: crawler access through Cloudflare's AI Crawl Control, Search Console's generative AI report, Bing Webmaster Tools citation data, prompt monitoring, GA4 segmented by AI referral source, and asking buyers how they found you. His line is the one worth stealing: a bot request is not a citation, a citation is not a visit, and a visit is not a business outcome. Convergence across signals is the proof, not any single number.

A related post cited a MarTech data study claiming 97% of AI-driven traffic is invisible in analytics. The summary: brands mentioned by ChatGPT or Perplexity see site visits climb to 1.5 to 2.5 times baseline over the following week, but only about 2.5% of those visits carry a trackable referral, and most land after day one so they get credited elsewhere. We're relaying this from the poster's summary, so treat the exact figures as directional until you've read the original. Another practitioner flagged a smaller trap: as ChatGPT adds paid placements, ad clicks and organic citation traffic can share a referral path in GA4, so untagged ads will contaminate your AEO numbers.

The other thread worth noting is about being chosen versus being cited. One post pointed to analyst Suganthan Mohanadasan's sample of 57 ChatGPT conversations: brands named in ChatGPT's own search query were cited in the final answer 68.9% of the time, versus 2.1% for brands that were only fetched by a live search. It's one account, weighted toward software, so it's a direction and not a controlled study. The same post cited Conductor's repeat-run testing, where brand overlap on identical prompts ranged from 40% at purchase intent to 63% at comparison intent. The queries closest to a buying decision were the least stable.

What the research shows this week

The biggest headline number came from AthenaHQ's State of AI Search report, published September 29. Across millions of responses from eight LLMs, the average brand appeared in 16.3% of answers while category leaders appeared in 56.5%, roughly 3.5 times more. Brand domains went uncited in 84% of responses, and question-focused content drove over half of all citations. It's a vendor report, so read it as a benchmark and not gospel, but the gap between leaders and everyone else is the part worth sitting with.

Search Engine Journal's September 30 piece, drawing on a Muck Rack analysis of more than 25 million AI-cited links, found Wikipedia is the most-cited domain in ChatGPT, and that about 42% of those Wikipedia citations point to pages about processes, technologies, or methods rather than brands or people. YouTube ranked third across major AI engines in Contently's 2026 analysis. The pattern across all of it: models lean on third-party, authoritative sources, which is why earned media and PR keep showing up as the underrated half of AI visibility.

On the Google side, Search Engine Roundtable's October 1 and 2 recaps note AI Overviews now appear for almost every large brand name query, with one dataset showing branded-query coverage jumping from 26% to 80%. Google is also testing tracking parameters on AI Overview links so publishers can finally separate that traffic, which would fix a real reporting gap if it ships. And its AI contribution pilot, which pays around 100 publishers through Search Console, reportedly works out to roughly 0.1% of ad revenue.

One older study is still the reference point for engine-by-engine reporting: Wellows' analysis of 22.7 million citations (published July 29, 2026, so background, not new) found 79.6% of cited sources appear on only one of five engines. If your visibility report blends ChatGPT, Gemini, and AI Overviews into one score, this week's measurement debate on LinkedIn is the reason to split it.

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