Brands are flooding AI search with new tactics to boost visibility. Some strategies are already backfiring. Here is what actually works and what is getting penalized.
Reddit’s presence in ChatGPT citations dropped sharply in August, falling from nearly 4% to just 0.5%. This shift came as AI models began to treat Reddit as a less reliable source, right when marketers were ramping up efforts to boost brand mentions on the platform. The pattern is familiar: as soon as brands find a way to game answer engine optimization (AEO), the models adjust and close the loophole.
The impact for publishers and brands is immediate. AI-driven search engines now decide which content appears in answers, and old SEO tactics are losing their effectiveness. Companies like Envive, Trendos, Profound, and Brandi AI have built businesses around helping brands appear more often-and more favorably-in AI search results. But even these firms admit that the logic behind large language models (LLMs) is still a mystery. No one can guarantee which content will be cited or why certain sources are chosen.
"In comparative tests, Reddit accounted for 17.87% of citations in ChatGPT, 9.07% in Perplexity, and 3.43% in Gemini, highlighting how AI engines weigh sources differently."
- Ranqo
Some trends are starting to show. Apparel brands see better results when their own sites use clear Q&A formats, which LLMs tend to treat as authoritative. Consumer finance brands, on the other hand, rely mostly on third-party citations like listicles and affiliate articles. Gintare Rimolaityte of Trendos points out that each industry needs its own approach, while James Cadwallader of Profound stresses the importance of fresh, regularly updated content. As LLMs rotate through sources, older material is quickly replaced, and only highly structured, comparison-based content tends to last.
Attempts to manipulate the system often backfire. The recent push to flood Reddit with brand mentions-sometimes called the "Reddit-ification" of AEO-has led to a crackdown from Reddit itself. The platform now uses its own AI tools to spot and suppress stealth marketing. A Reddit spokesperson has questioned the claims of any provider promising better LLM performance through Reddit posts. The steep drop in Reddit citations after ChatGPT’s August algorithm change shows how quickly these tactics can be shut down.
Promptwatch reports that Reddit’s share of visible ChatGPT Search citations fell from 3.83% between July 18 and August 7, 2026, to just 0.52% from August 14 to 17-an 86.4% decrease. A Kroma Studio analysis links this to a sudden change in source ranking after August 8, 2026. Prosperity Media’s review confirms that after this point, Reddit’s share in ChatGPT Search citations dropped below 1%, highlighting how volatile AI-driven rankings can be.
"Despite the sharp drop in visible citations, external reviews indicate that Reddit may still be used in background research by ChatGPT, even if it is not shown as a final cited link. This suggests that a decline in citation share does not necessarily mean Reddit is excluded from the model's research process."
- CyberNews, Editorial
With Reddit-focused tactics losing ground, AEO vendors are shifting to direct outreach. Companies like Noble automate requests to publishers, asking them to add client mentions to evergreen or affiliate content. Rahul Jain of Noble says these edits don’t compromise journalistic integrity, since the targeted articles aren’t traditional reporting. Noble’s system analyzes LLM citation data to decide which articles to target, sometimes leading to multiple clients competing for the same spot. The outreach is automated but designed to look personal, using different sender names and backgrounds.
Not all publishers are interested. Noble’s response rate is between 4% and 10%, and some clients have dropped the tactic after running into resistance. One ad tech executive left Noble’s program within two weeks, worried about brand reputation and publisher pushback. The risk of being seen as spammy or manipulative is real, and the payoff is uncertain.
Other AEO firms are watching closely. Leah Nurik of Brandi AI warns brands to avoid "tricky tricks" and focus on creating human-driven, AI-optimized content. She argues that trying to game the system is short-lived and damages consumer trust. Shafqat Islam of Optimizely agrees, advising brands to stick to the basics: make content that’s engaging, well-structured, and easy for both people and bots to understand.
What actually works
Despite the flood of AEO vendors and new tactics, the main principle hasn’t changed. Content that is genuinely useful, current, and clearly formatted still performs best in AI search. As previously reported, publishers who focus on content that survives AI search answers are more likely to keep their visibility and traffic as algorithms change.
Brands looking for shortcuts-whether through Reddit, publisher outreach, or other schemes-are seeing less and less return. AI models are getting better at spotting and penalizing manipulative tactics, much like Google did with SEO in the past. The lesson for marketers is simple: invest in content that serves real users, not just algorithms, or risk being pushed aside as AI search evolves.
As the AEO arms race heats up, the winners will be those who avoid chasing every new hack and instead focus on transparency, authenticity, and editorial rigor. Platforms are moving faster than marketers, and the cost of playing games is now measured in lost trust, wasted budget, and shrinking visibility. The only sustainable approach is to build content that earns its place, because the AI engines are watching-and they are not easily fooled.
