We read what a brand actually does, then propose the pillar pages and clusters that would make it the obvious source in its category. No search volume involved. Keyword tools recommend roughly the same things to everyone in a category. That is why so many content plans look alike and why so few of them build authority.
10 min read
By Alex Jordan, Founder
We give it one thing: a URL.
It reads the site the way a model would — what this brand actually does, what it sells, what territory it could plausibly own — and proposes three to five pillar pages, each with six to ten clusters underneath, ordered by what to ship first.
Every title has to work at the category level. No brand name allowed, anywhere in the plan.
That single constraint is what separates this from a content calendar. "Why [Brand]'s Merino Base Layers Last Longer" is a page about you, and an engine has no reason to cite it when someone asks about merino base layers. The category-level version is the one that gets lifted, and it is the one that earns you the citation.
LLM search engines cite the source that owns a topical territory. Not the one with the most branded pages.
No search volume. No SERP data. No embeddings. No competitor sitemaps.
That is not a gap, it is the design. Search volume tells you what people typed into a box last year. It does not tell you what a model will reach for when someone asks a question in their own words — and the whole point of this is planning for the second thing.
Depth gets cited. Coverage gets ignored.
This runs after the Pillar Analyzer, never before. Generating a plan without first grading what a site already has produces work that competes with the site's own pages.
The Pillar Cluster Engine is the sixth of the nine tools in our method. See all nine, in the order we use them →
The Pillar Cluster Engine powers two engines: Authority Cluster, building topical depth the models recognise, and Geo / Vertical Scale, where the same mapping generates the regional and vertical variations worth building pages for.
Your free AI Perception Report shows which territory you could own.
Firon's Pillar Cluster Engine is given one thing: a URL. It reads the site the way an AI model would, working out what the brand does, what it sells and what territory it could plausibly own. It then proposes three to five pillar pages, each with six to ten clusters underneath, ordered by what to ship first. It deliberately uses no search volume, no SERP data, no embeddings and no competitor sitemaps. Search volume tells you what people typed into a box last year. It does not tell you what a model will reach for when someone asks a question in their own words.
Because AI engines have no reason to cite a page about you when someone asks about the category. The Pillar Cluster Engine's hard rule is that every title works at the category level, with no brand name anywhere in the plan. "Why [Brand]'s Merino Base Layers Last Longer" is a page about the brand, and an engine has no reason to cite it when someone asks about merino base layers. The category-level version is the one that gets lifted and earns the citation. LLM search engines cite the source that owns a topical territory, not the one with the most branded pages.
No. In Firon's method the Pillar Cluster Engine runs after the Pillar Analyzer, never before. Generating a content plan without first grading the existing library produces a plan that competes with the site's own pages, splits authority across near-duplicates, and can bury the pillars that were actually working. The Pillar Analyzer grades which subjects the brand genuinely owns, which it touched and abandoned, and which have volume but no structure. The Cluster Engine then plans around that, so new work builds on existing depth instead of competing with it.
Not on its own. Depth gets cited and coverage gets ignored. LLM search engines cite the source that owns a topical territory, which is why the Pillar Cluster Engine plans three to five pillars with six to ten clusters each rather than a long list of disconnected posts. The step before it, the Pillar Analyzer, often recommends publishing less: consolidating four thin pieces into one and redirecting the rest so the pillar starts holding weight. The same mapping also powers Firon's Geo / Vertical Scale work, generating the regional and vertical variations worth building pages for.
Alex Jordan
Founder, Firon Marketing
15+ years scaling brands. A strategic partner for high-growth founders, focused on sustainable revenue models and long-term equity value.
