It doesn't matter what people ask ChatGPT

I’m not sure about you, but I am so tired of talking about AI search. But there’s just so much misinformation floating around, so many different products you could buy, so many salespeople. And no one wants to sit through another demo.

Do you even know what works for AI search citations? If not, that makes sense. Most of the process is opaque, and most of the SaaS products promising to show you how many times your institution or organization shows up in AI search answers are light on why that matters, if it matters at all.

So first off, if it’s a branded search — just like in SEO — you’re probably going to get cited or at least included in the research. Second, in most cases that person searching on ChatGPT or Google has used it before. They’re probably logged in. Just like when you use those apps for anything, it remembers context. So answers are going to be swayed by whatever other research they’re doing. Searches still don’t happen in a vacuum.

I said this on a live podcast in July 2026:

“It doesn’t really matter anymore what questions someone is asking ChatGPT — it matters what questions ChatGPT is asking your website after that happens. Because of how query fanout works, you can ask one question, but it’s broken up into 40, 50, 60 questions you never even see.”

That’s the real truth. AI search platforms don’t really answer that one question, and so if all you’re concerned with is whether or not you show up in ChatGPT answers, you’re putting too much weight on the wrong thing.

So what do you need to answer?

If the main question being asked in ChatGPT doesn’t matter… what does? It’s those 40+ additional questions that get asked under the surface. This is called query fan-out. When you type a question into your AI tool of choice, it parses that question with a model, generates a bunch of sub-questions, sends agents out to research them, weighs all the answers that come back, and then formulates the answer.

Basically, if you don’t show up in the answers to all of those sub-questions… you’re out of contention. So why are you spending money trying to track citations?

It’s vanity metrics all over again.

The new word-of-mouth

Okay, AI search citations aren’t nothing. But keep them in proportion. Did you know that Google still owns around 90% of all searches globally — and closer to 94% of US desktop search (as of this writing)? Did you know that giant daily usage number about ChatGPT is inflated and not at all indicative of actual users searching? ChatGPT is less important in August 2026 than you think it is, and Google is more important than you’re giving it credit for.

The platform where people are searching matters less than the effect this hyper-mediated format is having on behaviour. Remember when word-of-mouth was the gold standard for marketing reach? I miss those days. People don’t talk to people as much as they used to and they trust predictive models more and more each day.

AI search is the new word-of-mouth. And if you’re not part of the conversation, you won’t be recommended.

This is about content strategy, plain and simple. Until you can pay to show up in citations or be considered more favourably (that’s not how LLMs work, btw, but maybe in some sort of RAG approach… I digress), the best you can do is what you’ve been doing. Improve your content. Answer questions better. Make it make sense. Get a handle on consistency. Be hospitable.

PR plays a role in AI search, too

AI search has an equivalent to backlinks in SEO — and that’s mentions. This is all about trust signals. Do other websites, people, blogs, news sites talk about you? Do they link to you? How do they talk about you?

In higher education, that means the big brands — R1s, D1s, Ivies — are going to win out because they have the reputation, recognition, and online reach. That also means that if your small regional college has bad reviews online (like on Reddit or other discussion forums), you’re fighting an uphill battle. And if you’re never in the news, never ranked well in Google, and don’t talk about your distinct traditions much online… you might want to start thinking about that.

AI search is a content challenge

I have some good(ish) news. You don’t need to get in the weeds with how AI search works or spend loads on citation tracking. This is a content and brand problem, and you own both of those, hopefully. If you focus on creating a website that’s great for your visitors, you’re gonna be fine. If your product matches your brand promise, you’ll have fewer negative conversations happening outside your bubble.

But it doesn’t really end there. You need to be intentional about getting people to say the good things about your institution, too.

Friend, this isn’t new. Every business has dealt with this for decades. We were talking about this in my first college web job. And my second. And I’ve had this conversation with just about every client I’ve ever worked with. The systemic marketing issues in higher ed tend to get in the way of these actions.

The fix is to stop focusing on the metrics and the output until you’ve fixed the underlying issues. Make your website faster. Write your content for your audience, not leadership. Put better journeys in place. Personalize those journeys. Anticipate what your prospective student wants next.

You have the tools, I’m sure. But more likely than not, you don’t have them set up to give you a full picture of what your web visitors, prospective students, and accepted students need right this moment.

I’ll leave you with one hard truth. There’s no magical AI layer that can fix bad content and poor experiences. You’ve got to do the work to fix your marketing operations. Just like you’ve always needed to. AI puts a magnifying glass on what’s working and what’s not. Use that to your advantage and fix things.