The research is from Cornell. Thirteen words. One Reddit comment. That is all it takes to change what ChatGPT tells your prospect when they ask who to trust in your category. This is not speculation. Researchers tested it across hundreds of queries and found that a snippet as short as 13 words, appended to a Reddit post in a relevant subreddit, consistently changed the output of AI search tools.

The mechanism is simple. AI research agents scan user-generated content. Reddit accounts for roughly a quarter of all citations these agents produce. The way they decide what to cite is lexical similarity: they surface content that reads like what the user is searching for. Brands who figured this out first started writing Reddit comments that mirror the exact queries their customers type into AI tools.

Companies like RedRover exist specifically to do this. They charge to place branded content on Reddit with the stated goal of manipulating AI search results. This is not underground. They advertise it openly.

What this means if you run a B2B marketing function

If your company is absent from Reddit, Quora, and the other user-generated content platforms that AI tools scrape, you are absent from AI search. Not buried. Absent.

Your competitor does not need a sophisticated content operation to beat you here. They need 13 words and a relevant subreddit.

The field is not level. It never was. AI search made the imbalance cheaper to exploit.

What this means if you run your own outreach

You have probably noticed that ChatGPT and Perplexity now shape your prospects' research before they talk to you. When a prospect asks "best B2B email outreach tool for DACH market" or "why is my cold email failing," the answer they get is not neutral.

Someone is already trying to be in that answer. Either you are, or you are not.

Two moves

The defensive move: Find out what AI tools say about you right now. Open ChatGPT. Search for the exact problems your product solves, the category you compete in, and your company name. Read the answers. Note who gets mentioned and why. This is your baseline. You probably do not know it yet.

The offensive move: Authentic presence on the platforms AI tools cite. This does not mean spam. Moderation on Reddit is getting stricter specifically because AEO spam is overrunning subreddits. What it means: answer real questions in the communities where your buyers already are. Mention your tools when directly relevant, not every time. Write in the register of the community. Do not announce your product. Be useful. The comment that helps someone is also the comment that AI search cites.

The harder part

The Cornell researchers did not find a clean fix. Reddit and Wikipedia moderators are trying to keep up, but the volume of AEO content is increasing faster than human moderation can handle. AI companies are aware. Nothing has shipped yet.

In the meantime, the brands that show up in AI search answers are the ones that have a presence in the places AI search scrapes. This is not a moral position. It is a mechanical one.

Free tool
Find out if your copy would survive being scraped and evaluated.

The Bullshit Detector scores your copy on the same signals that make content credible or dismissable, whether the reader is human or a language model. Free. No signup.

Run the Bullshit Detector

Source: "Deep-research agents can be poisoned via user-generated content," Triedman, Zhang, Shmatikov. Cornell University, June 2026. Reported by 404 Media.