Lessons we learned the hard way
After applying almost every AI search best practice that researchers and experts said was required to appear in LLM (AI) answers.
↓
ChatGPT, Gemini, and Perplexity do not just mirror Google rankings. They look across your website and the rest of your online presence for signs that your business is clear, credible and worth recommending.
After applying almost every AI search best practice that researchers and experts said was required to appear in LLM (AI) answers.
↓
“AI does not leave a blank when it does not know a business. It fills the gap with whoever sounds closest and presents that answer as fact.”
In our tracking, early answers confused the company with similarly named businesses more often than they identified it correctly, mixing No Fluff up with No Fluff Jobs, No Fluff Agency, and No Fluff Selling, and delivering those wrong answers with complete confidence.
Through wrong meetings, misaligned proposals, and deals that never close because a buyer’s first impression was shaped by a confident, wrong answer they had no reason to question. During our experiment, we even received inquiries from people looking for one of our similarly named competitors, not realizing we weren’t who they thought we were.
“Media and publicity is not just awareness. It is how AI decides a business is credible enough to recommend.”
We tracked every source AI pulled when it cited or mentioned us across 13,000 AI answers over 90 days. Our website helped AI understand who we were when we were named in a question, but it was outside coverage that helped AI start recommending us in broader answers where our name was not part of the question.
The sources that help AI decide who to recommend are almost always outside the website: social media, publications, expert commentary, and credible third-party mentions. Public relations is no longer just awareness, it’s how AI decides a business is credible enough to recommend.
AI visibility is not stable, even when a business is gaining traction. When OpenAI updated its models March 3 through 5, our ChatGPT mentions dropped from 29 to 10 in one week, a 66% drop, while Perplexity moved from 46 to 47 and Google AI moved from 31 to 25. Same signals. Three completely different reactions.
The question every leadership team needs to answer is who owns this, who watched fluctuations and who coordinates across functions.
“The compounding we saw was not planned. It was a side effect of a precise marketing strategy.”
Instead of starting with a content calendar, we used those 150 questions to decide what needed to be on the site, what needed to be explained, and what needed to be prioritized.
The same specificity that made our content useful to AI made it useful to Google. The same questions we wrote for AI prompts became the exact phrases showing up in search queries we had never targeted, and the same thought leadership that earned AI citations drove LinkedIn engagement without a separate social strategy.
Focused strategy concentrates signal around a fixed set of answers until every channel that rewards relevance starts returning on the same investment. The compounding we saw was not planned. It was a side effect of precision.
The question is not whether AI search visibility matters. It is whether you know where your business stands. Let us build the same 150 buyer questions for you, run an audit, and give you a custom roadmap like we did for ourselves.