System 08
Content and SEO hubs
A content hub that is not written once and then left to rot, but moves daily with what happens inside your own systems. Stock, prices, lead times, recurring questions: those are figures only you hold, which is exactly why they are the only figures that set you apart.
Why AI content at scale usually returns nothing
The standard approach is familiar by now: have a model write a thousand articles about your subject, publish them, wait. The result is a thousand pages saying the same thing as your competitor’s thousand pages, because they were trained on the same sources. Google sees that, and since the helpful content updates such material largely drops out of view.
AI answer engines are stricter still. They cite a source only when it holds something that is not available elsewhere. A page that rephrases commonly known information never gets quoted, however well written it is.
The difference is not the model or the volume. It is whether there is data underneath that belongs to you.
How the system fits together
Internal data first, writing second, and a gate deciding what gets published at all.
- 01
Your own data as the source
Stock levels, lead times, price movement, return reasons, the questions that keep coming up in support. Facts from your own systems that exist nowhere else, and that give a page something to say.
- 02
Reacting daily to signals
When something meaningful changes, the page it concerns gets updated. A product back in stock, a lead time creeping up, a question suddenly asked often: that is the trigger to rewrite something, rather than waiting for the next content round.
- 03
Writing with an editorial step
The model proposes, a person approves. Facts are only carried over from fields you designated, so the model cannot invent anything at the point where invention does the most damage.
- 04
Measure and prune
Per page, whether it is indexed, pulls traffic and gets cited. Whatever does nothing after a few months comes out. That last step almost never happens, which is precisely why content programmes silt up.
How it gets built
- 01
Work out what data you hold
Which figures sit in your systems that nobody outside your company has, and are they factual enough to build a page on. If the answer is no, building makes no sense and I say so before money goes in.
- 02
Start small and steer on indexation
A subset of a few dozen pages. Only expand once those get indexed and pull traffic, because otherwise you publish something nobody sees in one go.
- 03
Switch on the daily cycle
Once the base holds, the connection to your systems goes live and the hub moves with it. From that point maintenance stops being a project and becomes a process.
What the system does
It produces pages that stay current without anyone maintaining them, and that assert things resting on your figures rather than on generalities.
- Pages built on facts from your own systems, not the same sources as your competitor
- Daily updates driven by what actually changes
- An editorial step in front of everything published
- Visibility of which pages work, and pruning of the ones that do not
What it costs
What an engagement costs depends on its scope and on how many systems are involved. I work with a fixed price per phase rather than open-ended billing: you know what a phase costs before it starts, and you decide after each one whether to continue.
I only quote a figure once I know what we are talking about. That takes a half-hour conversation, at no charge.
Whether this fits
This works well when
- your systems hold data nobody outside your company has
- your subject carries enough variation to justify several pages
- someone will keep reviewing the proposals
- you are willing to throw away pages that return nothing
This fits less well when
- the only information available also sits on a hundred other sites
- you want as many pages published as fast as possible
- nobody has time to edit before anything goes live
Questions about content hubs
Is this not just AI content, which Google penalises?
Google does not penalise AI, it penalises content with no value of its own. Their guidance says as much: how something was produced does not matter, whether it adds anything does. The difference sits in the data underneath. A page saying a product ships within two days because your stock system says so adds something. A page explaining what that product generally is does not.
How many pages does this produce?
Your data decides that, not your ambition. If each subject holds enough unique facts to fill a page that stands on its own, that can run to hundreds. If it does not, a threshold holds the page back. Better fifty pages that get cited than five hundred that get ignored.
Who actually writes it?
The model writes the proposal, someone on your side approves it. That is not a formality: in the first rounds things regularly go back because the tone is off or an assumption crept in that the data does not support. As the instructions sharpen that step gets lighter, but it does not disappear.
Does this work for AI answer engines?
That is what it is built for. ChatGPT, Perplexity and AI Overviews cite sources that say something unavailable elsewhere, and that do so in short, self-contained passages. Your own figures plus a clear passage structure is exactly what that takes. There are no guarantees, but without your own data the chance is zero.
Let's discuss how I can help your business grow
A short call is enough to tell whether there is a good fit. No sales pitch, just an honest read on what is possible.