Every article we produce starts from evidence — a strategy gap, a customer review, a Reddit thread, or a taxonomy hole. Never from a blank page. The worst content plans start with a blank page and a brainstorm. Ours start with evidence, and there are exactly four kinds.
9 min read
By Alex Jordan, Founder
Strategy-based — from the SEO and GEO strategy for the account. The gaps we already know about.
Review-based — from what customers actually said. Their language, their use cases, their objections.
Reddit-based — from live discussions where buyers are describing the problem in their own words.
Cluster-based — from the pillar map. The topical regions where we have nothing and a competitor has depth.
Every article traces back to one of those four. If a topic cannot be traced to evidence, it does not get written.
An article written from a keyword is written for a machine that no longer works that way. An article written from a real review answers a question a real buyer asked, in the words they asked it in — which is exactly the shape a model wants to lift and attribute.
The four modes are really four kinds of proof that someone wants this answered.
Everything routes into the client's own article tracker, so the plan, the draft and the status live in one place rather than in an email chain.
This is how we decide what to write today, run by our team using our own tooling. The fully automated version is in build. We would rather describe the method accurately than claim a robot writes your content — and if we did claim that, you should probably not hire us.
Your free AI Perception Report includes the first four topics we would write for you.
The Article Production Agent is the last of the nine tools in our method. See all nine, in the order we use them →
Every article traces back to one of four kinds of evidence, and if a topic cannot be traced to evidence, it does not get written. Strategy-based topics come from the SEO and GEO strategy for the account: the gaps already known. Review-based topics come from what customers actually said, in their language, with their use cases and objections. Reddit-based topics come from live discussions where buyers describe the problem in their own words. Cluster-based topics come from the pillar map: the topical regions where the brand has nothing and a competitor has depth.
Because an article written from a keyword is written for a machine that no longer works that way. An article written from a real customer review answers a question a real buyer asked, in the words they asked it in, which is exactly the shape an AI model wants to lift and attribute. Firon's four modes, strategy, reviews, Reddit and the pillar map, are really four kinds of proof that someone wants a question answered. A keyword alone proves only that someone typed something into a search box.
No. Firon's Article Production Agent describes how Firon decides what to write today, run by Firon's team using Firon's own tooling. A fully automated version is in build. Firon would rather describe the method accurately than claim a robot writes your content, and if an agency did claim that, you should probably not hire it. Everything routes into the client's own article tracker, so the plan, the draft and the status live in one place rather than in an email chain.
They come from Firon's review mining, which pulls verified reviews from the open web, including Reddit, Google and wherever customers actually talk, rather than only from a brand's connected reviews app. Review-based topics are usually the best article ideas in the account, because they carry the customers' own language, real use cases and real objections. That language is also the language buyer queries are phrased in, which makes an article built from it easier for an AI engine to match to a question and cite.
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.
