AI SEO Is Not a Thing

Sometime in early 2025, a switch flipped in the SEO industry. Overnight, every agency, every consultant, every tool vendor needed an "AI SEO" offering. LinkedIn filled with people who had been doing content audits six months earlier and were now "AI Search Strategists." Conferences added AI tracks. Newsletters pivoted. Agencies created new service tiers with new price points. The same twelve-person shops that couldn't staff a technical audit were suddenly experts in large language model optimization.

I watched this happen with the same feeling I get when I see a new restaurant open in a location where four restaurants have already failed. Admiration for the confidence. Certainty about the outcome.

Here is my position, stated plainly: there is no such thing as AI SEO. There are things that are changing about search, and some of those things are driven by AI. But the changes do not constitute a new discipline. They do not require new experts. They do not justify a new line item on your invoice. What they require is the same thing search has always required, which is that you be good at SEO, the regular kind, the kind that has been the same four things for twenty-five years.

I have an AI Search Optimization page on this very website. I will get to that.

The Same Index

When ChatGPT answers a question, it does one of two things. Either it draws on its training data - the vast corpus of text it absorbed before its knowledge cutoff - or it searches the web in real time via retrieval-augmented generation. Perplexity does the second thing more aggressively and more transparently. Gemini does both, because it is Google Search wearing a conversational skin.

The real-time search component queries a search index. The same kind of search index that Google has maintained since 1998. The same signals determine what surfaces: content quality, backlinks, domain authority, structured data, topical relevance. When Perplexity cites your page, it is almost always because your page already performed well in traditional search. The retrieval layer is not magic. It is search. It has always been search.

The training data component is different, and I want to be honest about that, because glossing over it is where most "AI SEO is just SEO" arguments get lazy. LLMs absorb Reddit threads, academic papers, forum posts, niche blogs - content that may never have ranked on page one of Google. The signals that determine whether a model "knows" about you are not identical to the signals that rank you for a keyword. They are closer to cultural presence than to search position. How often your name appears across the internet. In what contexts. Associated with what expertise. This is more like reputation than like optimization, and it is a real difference.

But here is the thing: that difference does not create a new discipline. It creates an additional consideration within the existing discipline. The way mobile-first indexing created an additional consideration. The way HTTPS migration created an additional consideration. The way internationalization creates additional considerations. None of those spawned permanent sub-specialties. They were absorbed into what we already did, because that is what competent practitioners do - they absorb new developments instead of pretending each one requires a new certification.

The Measurement Problem

Here is the strongest argument for AI SEO being its own thing, and I want to give it its full weight because I take it seriously.

SEO, as a practice, is built on a feedback loop: change something, measure the result, iterate. The entire methodology depends on measurement. And in AI search, the measurement infrastructure does not exist. There is no Search Console for LLMs. There is no rank tracker for ChatGPT. Ahrefs and Semrush have built tools that claim to track AI mentions, but they are sampling from a stochastic system - the same query produces different citations at different times for different users. You cannot measure what you cannot reproduce.

This means the core methodology of SEO - the test-measure-iterate loop that has defined the practice since its inception - breaks in the AI context. And if the methodology breaks, doesn't that imply a different discipline is needed?

No. It implies the methodology has a gap. A gap is not a discipline. You do not build a new field of medicine because a new disease resists existing diagnostics. You extend the diagnostics. The measurement tools will come. They are already being built, badly, and they will get better. When they arrive, they will be integrated into the same dashboards and the same workflows and the same monthly reports. They will not be a separate practice. They will be a new tab in Search Console. And the people who spent three years calling themselves "AI SEO Specialists" will quietly update their titles, the way the "Voice Search Optimization Experts" of 2018 did, the way the "AMP Specialists" of 2017 did, the way every manufactured sub-specialty in this industry eventually does.

The Zero-Click Difference

There is another real difference and I am not going to minimize it.

In traditional search, Google sends you the click. You rank, someone clicks, they land on your site, you convert them. The whole value chain depends on that click. In AI search, the model often takes your content, synthesizes it into an answer, and delivers it to the user without them ever visiting your site. Your content was used. You got nothing. No traffic. No conversion opportunity. No data.

This is genuinely different. This is not "a new distribution channel." This is your distribution channel being consumed. The economics are not slightly adjusted. They are, in certain query categories, inverted.

I take this seriously. Anyone who dismisses it is not paying attention. But I want to point out that this is not an optimization problem. You cannot "optimize" your way into getting credit for content that an AI has already synthesized into its own response. This is a business model problem, a legal problem, a platform economics problem. Calling it "AI SEO" and hiring a consultant to handle it is like hiring a plumber to fix your lease terms. The skill set does not match the problem.

The people who will navigate the zero-click shift successfully are not the people with the best "AI SEO frameworks." They are the people who build brands strong enough that the AI has to mention them by name, because not mentioning them would make the response obviously incomplete. That is brand strategy. That is SEO in the deepest, oldest sense of the word - making yourself so authoritative that the systems cannot ignore you, regardless of what form those systems take.

The Industry Playbook

I have been in this industry since 2000. I have watched this exact cycle play out at least six times. A new technology arrives. It is genuinely important. And then:

The technology gets an acronym. The acronym gets a conference track. The conference track gets speakers. The speakers get consulting practices. The consulting practices get case studies. The case studies get challenged. The challenges get ignored. The invoices get sent. The technology gets absorbed into regular SEO. The specialists quietly become generalists again. The acronym fades. Nobody talks about it at the next conference because there is a newer acronym by then.

Voice search optimization. AMP. Mobile-first SEO. E-E-A-T optimization. Core Web Vitals consulting. GEO. Now AI SEO.

Voice search was supposed to be the future of search. By 2020, 50% of all searches would be voice, they said. It turned out to be writing clear answers to questions, which was just content strategy, which was just SEO. The specialists evaporated. The agencies quietly removed the service page.

Every single time, the people who said "this is just SEO" were right. Not because nothing changed. Things changed. But the changes were absorbed into the existing practice, because that is what practices do. Medicine did not split into "antibiotic medicine" and "non-antibiotic medicine" when antibiotics arrived. It absorbed antibiotics into medicine. The practitioners who understood this adapted. The ones who built their identity around the new thing got stranded when the new thing became the normal thing.

But What If I Am the Print Journalist

There is a version of this story where I am wrong. I want to tell it to you because I owe you that.

In 1998, plenty of print journalists said the internet was just a new distribution channel for the same journalism. The writing was the same. The reporting was the same. The skills were the same. They were right about all of that. And many of them lost their careers anyway, because the distribution shift was large enough to restructure the entire industry, and "the fundamentals haven't changed" turned out to be a correct observation that led to the wrong strategy.

I am forty-two years old. My competitive advantage is twenty years of pattern recognition built on Google's paradigm. If that paradigm shifts enough, that pattern recognition becomes muscle memory for a game that is being replaced. Someone who has spent two years studying how LLMs select and surface information might understand the emerging landscape better than I do. The PE partners and startup founders I work with are increasingly asking about AI search, and if my answer is always "it's just SEO," there is a version of the future where that makes me the smartest person in the room and another version where it makes me the guy who was right about the music but wrong about the business.

I hold both of these possibilities in my head simultaneously. I think I am right. I think the foundation is the foundation. But I have been wrong before, and the times I was wrong were usually the times I was most confident, and I notice that I am quite confident right now, and that makes me nervous.

The honest answer is: I do not know for certain. Nobody does. The people selling certainty on either side - "AI SEO is the future" or "nothing has changed" - are both overstating their case. The difference is that my uncertainty costs you nothing, and their certainty costs you a retainer.

What I Actually Tell Clients

When someone asks me about AI SEO, here is what I say. It is not exciting. It does not fill a webinar.

Do the work that is robust across multiple possible futures. Fast site. Clear content that genuinely answers questions. Real authority built through real expertise. A brand that people reference by name. Clean technical foundation. Internal links that make sense. These things work for Google. They work for AI search. They work for whatever comes after AI search. They have worked, in various forms, since I started doing this in 2000, and they will work after I stop.

Pay attention to where your traffic is actually coming from. If you see AI referral traffic growing, study it. Understand which pages get cited and why. Adapt. But adapt within your existing practice, not by bolting on a new one.

Do not pay for certainty that nobody has. The measurement tools do not exist yet. The case studies are anecdotal. The frameworks are speculative. Anyone selling you a guaranteed AI search strategy in 2026 is selling you confidence, not competence. I wrote about this in more detail in my piece on GEO, and everything I said there still applies.

If your SEO fundamentals are solid, you are already doing the work that drives AI visibility. If your fundamentals are not solid, no amount of "AI optimization" will compensate. The foundation is the foundation. The new systems are built on the old systems. Get the old systems right.

The Page on My Website

I have an AI Search Optimization page on this website. I said I would get to that, and here I am.

I keep it there because when a potential client searches for AI SEO services, I want them to find me instead of someone who will charge them ten thousand dollars a month for a cargo cult. This is the game. I am in the game. I am playing it while telling you the game is partially invented, which makes me a hypocrite, which I have been before and will be again, and which I prefer to being a hypocrite who doesn't admit it.

When someone hires me through that page, what they get is SEO. The boring, foundational, twenty-five-year-old kind. Their AI visibility improves as a side effect, because that is how it works, because the systems are built on the same foundation, because there is no shortcut around the foundation, because there has never been a shortcut around the foundation.

I would rather give you the correct answer than the billable one. The correct answer is: there is no AI SEO. There is SEO, done well or done poorly, and the AI systems can tell the difference the same way Google can, because they are, at bottom, reading the same internet you built your site on.

The interface changes. The foundation does not. It never has.

The interface changes. The foundation does not. If your site is fast, clear, authoritative, and answers real questions, the machines will find you. They always have. The only thing that changes is which machines.