Every emerging discipline eventually produces one organisation that everyone else references. Not necessarily the first mover. Not the biggest. The one that named things clearly enough, early enough, that their vocabulary stuck.
SEO had Moz. For almost a decade, if you wanted to understand how search worked, you read the Moz Beginner’s Guide. If you wanted to follow where the industry was going, you watched Whiteboard Friday. Rand Fishkin didn’t just teach SEO — he defined what SEO was, and the industry adopted his definitions. Domain Authority. Page Authority. The concept that links were votes. The framing that technical, content, and off-page factors were the three pillars. Moz didn’t invent all of these ideas, but they named them in ways that made them teachable, citable, and repeatable.
That is a different kind of authority than ranking well. It is the authority of having shaped how a field thinks about itself.
The Moz Lesson: Why the Company That Names Things Wins
Moz’s dominance between roughly 2007 and 2015 was built on a specific flywheel that most people remember but fewer people understand precisely.
The education came first. The Beginner’s Guide to SEO answered the question every newcomer was asking: what is this, and where do I start? It was comprehensive, clear, and freely available. Every person who found that guide and found it useful associated Moz with the answer. When they later needed tools, they bought Moz tools. When they advanced enough to attend conferences, they attended MozCon. The community formed around the education, not the other way around.
What Moz built, without necessarily framing it this way at the time, was a vocabulary franchise. They owned the language the industry used to describe itself. When a client asked “what’s your DA?” they were using Moz’s metric. When someone talked about “on-page, off-page, and technical” as the three pillars, they were using Moz’s framework. That vocabulary franchise produced extraordinary compounding authority because every conversation in the industry referenced Moz’s mental models, whether or not Moz was in the room.
The decline is instructive precisely because it wasn’t a collapse. Moz didn’t do anything catastrophically wrong. What happened was subtler: they shifted their focus from defining the field to competing in it. The Beginner’s Guide stopped being updated at the same pace the industry moved. Whiteboard Friday continued but became one voice among many as LinkedIn, Twitter, and YouTube democratised thought leadership. The tool suite faced direct competition from Ahrefs and SEMrush — both of whom competed not on education but on data infrastructure, and both of whom won that competition decisively.
The critical mistake wasn’t building tools. It was stopping the thing that made the tools secondary: being the place where the industry’s mental models got formed. Ahrefs didn’t beat Moz on education. They beat them on backlink data. SEMrush didn’t beat them on frameworks. They beat them on keyword volume. Moz’s actual moat was never the tools — it was the vocabulary. And when they stopped adding to the vocabulary, the moat drained.
The lesson: you do not defend a naming position by competing on features. You defend it by continuing to name things.
The AI Search Window Is Open — And It Won’t Stay Open Long
In March 2026, I ran a search for “AI Agent Optimisation (AAO)” and took a screenshot. The featured snippet was from seostrategy.co.uk. Below it was a Search Engine Land piece from February 2026. Below that was a LinkedIn article from Nathan Furr, posted a year earlier, framing AAO as a future layer on top of existing SEO efforts.
That screenshot is interesting not because of the ranking. It is interesting because the vocabulary is still being contested. The term AAO exists in three different places with three different framings, and none of them has definitively won. The field does not yet have its Moz moment — the point at which one source becomes the canonical reference that everyone else defers to.
That window is open right now, in 2026, in a way that it was not open for traditional SEO by the time most practitioners arrived. The vocabulary of AI search is forming in real time. Terms like GEO, AEO, AIO, AAO, and AI visibility are being defined and redefined across blog posts, LinkedIn articles, conference talks, and academic papers. The definitions are inconsistent. The frameworks are competing. The canonical reference does not yet exist.
This is the same moment Rand Fishkin was in when he published the first version of the Beginner’s Guide. The difference is that in 2007 there was no SEO authority. In 2026 there is no AI search authority. The conditions are structurally identical.
What I have been building at SEO Strategy Ltd over the past two years is an attempt to occupy that position before it closes. The AI Discovery Stack — a five-layer diagnostic framework mapping how AI systems progress from entity recognition to agentic action — was published in March 2026 with a specific provenance intention: to be the named framework that practitioners reference when diagnosing AI visibility failures. The 3Cs framework — Code, Content, Contextual Linking — developed in 2010, has been extended with an AI-era version that maps onto the Algorithmic Trinity that underpins every AI search platform. The OARCAS Framework was published the same week, with full scoring rubrics and version history, because frameworks without methodology are opinions.
The provenance matters. Moz’s authority compounded because every practitioner who came after them was learning a vocabulary that already existed, already had a source, already had an owner. The practitioner who publishes a framework in 2024 and watches it get cited without attribution in 2027 has created a public good, not a competitive advantage. The practitioner who publishes a named, dated, attributed framework and builds the entity chain — author, organisation, canonical URL, external citations — creates something that compounds differently.
I am not the only person working in this space. Jason Barnard at Kalicube has done foundational work on entity SEO and coined the Algorithmic Trinity framework. Search Engine Land is actively covering AI search methodology. Researchers at Princeton, Georgia Tech, and IIT Delhi published the GEO-Bench study that produced the 30–40% citation improvement figure I cite regularly. The vocabulary is being built by multiple people simultaneously.
But the Moz lesson is clear: the window for defining a field’s mental models is specific and time-limited. It does not stay open indefinitely. The vocabulary that sticks is the vocabulary that gets named clearly, early, and then reinforced continuously — through frameworks, through evidence, through the kind of practitioner authority that cannot be generated retroactively.
In 2007 the question was: who will explain SEO well enough that the industry adopts their vocabulary?
In 2026 the question is the same. The field is different. The window is the same.
Sean Mullins is founder of SEO Strategy Ltd. He has been building websites since 2005 and working in SEO since 2006. The AI Discovery Stack, OARCAS Framework, and 3Cs Framework are published at seostrategy.co.uk with full provenance documentation.
