The AEO Panic Is Overblown, and ChatGPT's New Index Proves It

There is a line I have heard in what feels like every strategy meeting for the past two years. AI search changes everything. Throw out the old playbook. The rules are gone.

I want to push on that, because I think it is mostly wrong, and this week gave me a clean example of why. The same week people were once again telling me that classic search is dead, we got the news that ChatGPT has been quietly building and leaning on its own search index. Google refiled a legal complaint over who gets to scrape and use search results data. And Google Analytics rolled out campaign benchmarking so you can compare your numbers to everyone else's.

Three separate stories. One quiet lesson underneath all of them.

The tools got new names. The physics did not.

What everyone is repeating right now

The going belief goes something like this. Search used to be about ten blue links. Now people ask a chatbot a question, get one answer, and never click anything. So SEO is finished, and a shiny new discipline called AEO, answer engine optimization, has taken its place. Different rules. Different mechanics. Different everything.

I get why this is a seductive story. It sells conference tickets. It sells consulting. It gives everyone a reason to panic productively, which is a favorite pastime in this industry. When something feels new, we all want to be the first person in the room who understands it.

But I have sat with enough clients through enough hype cycles to have developed a reflex. When someone tells me the rules are gone, my first question is simple. Okay, compared to what? What specifically is different, and what is just the same old thing wearing a new hoodie?

ChatGPT built an index. Read that again.

Here is the detail that I think matters most, and that almost nobody is sitting with long enough. The most talked about AI answer engine in the world needed to build an index.

Think about what that means. To answer your questions with anything close to current, trustworthy information, the model cannot just rely on what it absorbed during training. It has to go out, crawl the web, store what it finds in an organized way, and then decide which sources are worth pulling from when someone asks something. Crawl. Store. Rank. Retrieve.

If that sequence sounds familiar, it should. That is what a search engine has always done. Google did not invent magic in 1998. It invented a very good way to crawl pages, index them, and rank which ones deserved to be shown. ChatGPT arriving at the same basic architecture is not a plot twist. It is a confirmation.

When the future reinvents the past to function, that tells you the past was never the problem.

So when someone says the index era is over, I would gently point out that one of the biggest AI companies on earth just spent real money and engineering talent building an index. Not because it is nostalgic. Because you cannot skip that step. Somebody, somewhere, still has to have a good, organized, trusted map of what exists on the web and which parts of it are any good.

What actually changed, and what did not

I am not one of those people who pretends nothing is different. Something real did shift, and I want to be honest about it. The interface changed. The number of clicks changed. A person who used to scan a page of results and pick one might now read a synthesized answer and stop there. That is a genuine change in behavior, and if your whole business depended on being the fourth link people clicked out of curiosity, you should feel that.

That matters.

But notice what did not change. The machine still has to decide who to trust. It still has to find sources. It still favors content that is clear, well structured, specific, and corroborated in more than one place. It still struggles with thin, vague, me too pages that say nothing a hundred other pages have not already said. The thing that makes an AI recommend you is not some secret AEO incantation. It is the same thing that made Google rank you. Be a credible, specific, genuinely useful source that other credible sources seem to agree with.

I have watched this play out with our own clients. The pages that show up inside AI answers are, almost boringly, the pages that were already doing the fundamentals well. Clear expertise. Real specifics. Actual points of view instead of hedged mush. The businesses that got recommended by a chatbot did not do it by chasing a new acronym. They did it by being worth recommending. The AI is just a new referee applying an old standard more strictly.

So the honest version of the advice is not throw out your playbook. It is do the boring parts of your playbook better than you have been getting away with. AI is less forgiving of filler than a bored human with a mouse ever was.

The benchmark trap

Which brings me to the second piece of news, Google Analytics adding benchmarking so you can compare your campaign performance against other businesses. On its face this sounds helpful, and in some doses it is. But I want to wave a flag here, because benchmarks are one of the most quietly dangerous comforts in marketing.

Here is the trap. A benchmark answers the question how am I doing compared to others. It does not answer the question that actually pays your bills, which is am I doing well enough to hit my goals. Those are not the same question. They are not even close.

Being average in a weak field is not an achievement. It is a warning with good manners.

Say you find out your email open rate sits right at the industry average. Comforting, right? Except the industry average includes thousands of businesses sending lazy, poorly targeted email to lists they barely maintain. Matching them tells you almost nothing about whether your email is helping you grow. You have measured yourself against mediocrity and felt reassured by the tie.

I have seen founders make genuinely bad calls off benchmark data. They see they are above average on some metric and stop pushing on the exact thing that could have doubled their results. Or they see they are below average and panic and gut a channel that was actually working fine for their specific model. A benchmark is a hammer, and a hammer is not a house. It is one tool for one job, and the job is context, not conclusions.

Use the new benchmarks the way you would use a stranger's opinion. Interesting. Occasionally useful. Never a substitute for knowing your own numbers, your own margins, and what a customer is actually worth to you. The moment you outsource your definition of good to a global average, you have handed your strategy to a crowd of people who are not trying to run your business.

The Google fight nobody's really talking about

The third story is the one that looks like inside baseball but might matter more than the other two combined. Google refiled its legal complaint with new terms in a dispute over scraping and using its search results data.

Strip away the legal language and here is the real question underneath it. Who owns the data about what the web contains and how it ranks? Because whoever controls that index, that map of the web, controls an enormous amount of leverage over how everyone else finds anything.

Go back to the first story for a second. ChatGPT needed an index. Building and maintaining a fresh, high quality index of the entire web is brutally expensive and hard. So there is now enormous commercial pressure over who gets access to that data, who can scrape it, who can resell it, and who has to build their own from scratch. The AI answer era did not make the index less valuable. It made the index the single most contested asset in the room.

This is the part I wish more marketers would sit with. The narrative says AI freed us from search engines. The legal reality says the biggest players are fighting harder than ever over the exact infrastructure that search engines are built on. Follow the lawsuits, not the keynotes. Lawyers do not spend money fighting over things that stopped mattering.

So what do you actually do about it

Let me be practical, because I do not want this to read as smug from the sidelines. Here is how I am actually advising people, and it is unglamorous on purpose.

• Keep doing the fundamentals, and raise the bar on them. Clear structure, real expertise, specific claims, honest points of view. The things that got you ranked are the things that get you recommended.

• Assume the machine reading your site is smarter and less patient than the human it replaced. Vague, padded, say nothing content that used to scrape by will now get skipped entirely.

• Treat benchmarks as a thermometer, not a target. Know what a customer is worth to you and work backward from that, not from what the average company down the street is pulling off.

• Watch who controls the underlying data, because that is where your access and your costs will eventually be decided, whether you follow the news or not.

Notice what is not on that list. Panic. Rebranding your whole strategy around a three letter acronym. Firing your SEO person to hire an AEO person who does the identical job with a fresher business card.

The past keeps showing up because it works

Every few years this industry convinces itself that the ground rules have been rewritten. Social was going to kill search. Voice was going to kill typing. Now AI is going to kill clicks. And every time, the useful truth turns out to be less dramatic and more demanding than the headline. The interface changes. The channel changes. The plumbing stays remarkably the same, and the businesses that win are the ones that were already good and got a little more disciplined.

ChatGPT building an index is not a warning that everything is different. It is quiet proof that the thing we have been doing all along still had to exist. Somebody still has to crawl the web, judge it, and decide who is worth trusting. Your job has not changed. Be the source worth trusting, so clearly that whoever is doing the judging, human or machine, cannot reasonably leave you out.

The rules did not disappear. They got a new enforcer, and it reads more carefully than the last one.