Last month I did something that has now become a standing habit before every client strategy call. I opened an AI assistant and typed a question the way a real buyer types it. Not a keyword. A question.

"I’m looking for a good [your category] in Las Vegas. Who should I consider?"

Back came a short, confident, friendly paragraph naming three businesses, with a sentence about why each might be a fit. My client was not one of them. Two of their direct competitors were.

Nothing had gone wrong with my client’s marketing. Their website is beautiful. Their Instagram is active and well produced. Their SEO consultant had them ranking respectably on Google. None of that decided this particular answer, because this answer was not assembled the way a page of search results is assembled.

That was the moment this stopped being a trend piece to me and became a client conversation.

The behavior shift already happened

We are past the speculation stage. Pew Research Center found that about half of U.S. adults now use AI chatbots, up from a third two years earlier, with roughly a quarter using them daily. The single most common reason people give is to search for information.

That is the number that matters, because searching for information is exactly what someone is doing in the days before they spend money with you. Which contractor. Which agency. Which brand. Which studio. The bigger the purchase and the more trust it requires, the more homework people do first, and a growing share of that homework now happens in a chat window instead of a results page.

Here is what makes that different from a Google search. Google hands someone ten links and lets them judge. An AI assistant hands them one synthesized answer, written in a calm and confident voice, usually naming three or four options. There is no page two. There is no scrolling past the ads to find the organic results. You are either in the paragraph or you are not.

The part that should get your attention

So where does that paragraph come from?

Muck Rack’s Generative Pulse team studied more than 25 million links cited across ChatGPT, Claude, and Gemini and found that 84 percent of AI citations came from earned media, meaning journalism, academic and government sources, reference sites, and third party coverage. Journalism alone accounted for more than a quarter of everything cited. Paid and advertorial content accounted for three tenths of one percent.

Read that ratio again. Earned media at 84 percent. Paid content at 0.3 percent.

And the pattern has held steady across three editions of that study going back to mid 2025, which tells me this is not a quirk of one model release. It is how these systems are built to behave. They are trained to weight sources that appear credible and independent, and they treat your own promotional copy roughly the way a skeptical person treats it, which is to say they discount it.

Your website tells the machine what you claim. Everything else tells the machine whether to believe you.

This is the oldest lesson in PR wearing new clothes

I have spent my career on the relationship side of business, from healthcare technology rollouts to years working in Japan to two decades in professional networking. And when I first read those numbers, my reaction was not that something new had arrived. It was recognition.

Think about how a referral actually works. Someone asks a friend, "who did you use?" The friend does not read them your brochure. The friend repeats what she has heard, what she experienced, what a colleague mentioned, what she read in a publication she trusts. She is synthesizing third party evidence into a single confident sentence.

That is exactly what the AI is doing. It is behaving like a friend who has read everything and has no particular loyalty to you.

Which means the strategy has not actually changed. What other people say about you was always more valuable than what you say about yourself. The only thing that changed is that the gap is now measurable, and it now shows up in a paragraph a stranger reads at eleven at night before deciding who to call in the morning.

Run this audit today. It takes ten minutes.

I want you to do this before you do anything else, because opinions about AI are cheap and your own results are not.

Ask the question a buyer would ask. Open ChatGPT, Gemini, and one more assistant. Type the question the way a first time customer would type it, with your category, your city, and the actual problem they are trying to solve. Not your business name.

Write down who appears. Note every business named across all three tools. If you appear, note what the assistant says about you and whether it is accurate and current.

Ask a second, harder question. Try "what should I look for when choosing a provider for this," then see whose expertise gets referenced in the answer. That is the credibility layer, and it is often where you can win before you can win the direct question.

Look at what the answers are built from.. When the tool shows sources, look at what kind of source they are. You will usually see press coverage, industry publications, review platforms, and directories. You will rarely see anyone’s homepage.

Whatever you find, do not panic and do not celebrate. This is a snapshot, and these answers vary by phrasing and by day. What matters is the pattern.

What actually moves this

Get named by other people. Every credible third party mention is now doing double duty, building trust with humans and feeding the systems that summarize you. Say yes to the podcast. Give the trade publication the quote. Contribute the expert paragraph to the association newsletter. This is where our nearly thirty years of media relationships does work that no tool substitutes for, because getting into respected publications is a relationship business before it is a technology one.

Be quotable, not just visible. Assistants pull specific, attributable expertise. A founder who is consistently quoted explaining how something in their field actually works becomes the reference point for that category. Vague brand language gives these systems nothing to grab.

Fix your consistency. Same business name, same address, same credentials, same services everywhere you appear. These systems are pattern matching across sources. Contradictions get you discounted or skipped.

Keep reviews current and answered. Review platforms and directories show up constantly in these answers. Volume, recency, and how you respond all become part of the summary a stranger reads.

Stop measuring what does not matter here. Ad spend does not buy you into that paragraph. As the citation data shows, paid content is essentially invisible to it.

What not to do

Do not try to trick it. Every few months someone sells a shortcut for gaming AI answers, and every few months the models update and the shortcut stops working. Meanwhile, real coverage in real publications keeps compounding.

Do not flood the internet with AI written content about yourself either. Volume of self published material is not what these systems weight. Being talked about is.

The uncomfortable good news

If you are not in those answers today, you have not fallen behind on some technical standard you never heard of. You have a credibility gap, and credibility gaps are fixable. They are fixed the same slow, dependable way they always were, by doing work worth talking about and building relationships with the people who do the talking.

The machine did not change the rules. It just started reading your reputation out loud to strangers.

Go run the ten minute audit. Then let’s talk about what you find.

Want to know what AI is telling your prospective clients about you, and what to do about it? At JNR Communications, we combine Jennifer Rice's nearly 30 years of media relationships with my strategic networking expertise to help businesses earn the third party credibility that drives real results. Let's talk about your goals.

📞 (818) 357-7286 | ✉️ Greg@JNRcommunications.com