SEO content writing, built from what your business knows.

SEO content writing fails in two familiar ways: generic articles that rank for nothing, and good articles nobody can produce twice. We build the writing workflow that sits between them — researched against real searches, drafted from your own knowledge base, edited to a written standard, and run by your team after we hand it over. Every article answers its question first, which is what search engines and AI answers both look for.

What makes an article worth ranking.

Search engines and AI answers both reward the same thing: the clearest, most useful answer to the question someone actually asked.

That sounds obvious and is rarely done. Most SEO article writing buries the answer under an introduction, pads to a word count, and says what every other article says because it was written from the same search results. AI-assisted SEO writing made that faster, and made the sameness worse.

An article worth ranking answers first, then expands. It says something the business genuinely knows: an example from real work, a position it is willing to defend, a mistake it has watched clients make. That is the material only you have, and it is why we write from your knowledge base rather than from the search results.

The same article now has a second reader. AI answers lift single passages out of a page and quote them, so a section that only makes sense after the paragraph above it is a section nobody will quote. Writing for search and writing to be cited turn out to be the same discipline.

The anatomy of an article that ranks and gets quoted.

Every article the workflow produces is built to the same six rules, and the editing rubric checks each one.

The answer comes first

The opening sentence of the article, and of every section, answers its question. A reader skimming, or a model lifting one passage, gets the answer without needing the paragraph above it.

One question per heading

Each heading names one thing, and the section under it answers only that. It is what lets a search engine or an answer engine quote a section on its own.

Something only you know

An example from real work, a trade-off nobody mentions, a mistake you have seen made. Google’s own guidance asks whether content shows first-hand experience; this is where it comes from.

Evidence, or words

Every number has a source a reader can check. A claim we can’t source is said in words — often, rarely, most of the time — and never dressed up as an invented statistic.

A named author

A byline and a short note on who wrote it and why they know, because readers ask that question and Google tells writers to answer it.

Links that lead somewhere

Up to the page the article supports, across to related articles, and out to the sources it names. An article that links to nothing, and that nothing links to, is an island.

The writing workflow.

Every article passes through the same steps, so quality stops depending on who wrote it.

01

Choose the question

One search, one article. The question comes from the keyword map, checked so no two pages on your site compete for it, and sorted by intent: someone learning gets an article, someone ready to buy gets a service page.

Deliverable: a brief with the target question and what the article must cover
02

Research what exists

What already ranks, what it misses, and which claims need a source. Anything we can’t source is said in words, never with an invented number.

Deliverable: a research note with sources attached
03

Draft from your knowledge

Claude Code drafts against the brief, your knowledge base and your voice reference, with the answer in the opening lines of every section and the first-hand material marked for the editor to check.

Deliverable: a structured draft ready for editing
04

Edit to a standard

A written rubric covers accuracy, voice, the six rules above and the on-page basics. A named editor approves; nothing ships on the model’s say-so.

Deliverable: an editing rubric and an approval route
05

Publish, link and revisit

The article goes live with its links up to the page it supports and across to related articles, then joins the measurement loop. An article that falls behind the question it answers is refreshed, not replaced.

Deliverable: published articles, an internal-linking record and a refresh list

What Google has said about written content.

Much of the advice about writing for search is folklore. These are Google’s own published positions, and what the workflow does with each.

Question Google’s position What the workflow does
Is AI-assisted writing allowed? Using AI is not against Google’s guidelines; using it mainly to manipulate rankings is Uses AI to draft from your knowledge, with a person editing and approving every piece
What about publishing at volume? Many pages made mainly to rank rather than to help people is scaled content abuse, however they were produced Publishes one article per real question, at the pace good editing allows
Is there an ideal word count? No — Google says it has no preferred word count Writes until the question is answered, then stops
Does it matter who wrote it? Google asks writers to make clear who created the content and, where readers would wonder, how Puts a named author on every article, with a note on why they know
What does Google want instead? Helpful, reliable content written for people first, showing first-hand experience where it applies Builds every article on something the business has actually done or seen

Our own blog is written this way.

We don’t ask you to take the standard on trust. Every article on this site’s blog is written to it, and checked before it publishes.

An automated audit runs over every post before it goes out. It confirms the post targets one question that no other page here already claims, and that it links up to the service page it supports. It checks that every outside source the post names is linked, that the author named on the page is the author named in its markup, and that its FAQ matches its structured data word for word. A post that fails doesn’t publish.

None of that makes a post good — the editing does that. It makes sure a good post is never let down by something nobody checked, and it is the same discipline we build into a client’s writing workflow.

Three ways to get SEO content written.

Each one costs something different. The question is which cost you would rather pay.

A generic AI article A freelance writer A writing system
Knows your business No After briefing From the knowledge base
Sounds like you Rarely Sometimes By design
Has first-hand material Never If you supply it Drawn from your own work
Consistent quality Consistently average Depends on the writer Held by the rubric
Repeatable without you Yes, badly Only while they stay Yes, by your team

The honest cost.

Good search articles take editing time, and we will not pretend a model removes it. The workflow makes editing faster and more predictable; a person still reads every piece before it goes out.

It also needs a business with something to say. The first-hand material has to come from somewhere, which means someone who does the work giving it a little time each week. Without that, the workflow produces tidy articles that sound like everyone else’s.

And search is slow. Articles published now are the ones that rank later, and a new site waits longer than an established one. This is the right thing to build, but it should sit beside a channel that produces conversations sooner.

Questions people ask before they book.

Do you write our articles for us?

We write the first articles with you while the workflow is built, so the standard is set on real work. The deliverable is the workflow itself, with your team trained to run it. We do not sell an open-ended writing retainer.

Is the writing done by AI?

The drafting is AI-assisted, using your own knowledge base and voice reference rather than a blank prompt. The research, the choice of question, the first-hand material and the final edit involve people, and a named editor approves every piece.

Does Google penalise AI-written articles?

Not for being AI-written. Google’s published position is that using AI isn’t against its guidelines; what it acts against is content made mainly to manipulate rankings, including many pages published at scale that don’t help anyone, however they were produced. An AI-assisted article that answers a real question from real experience, edited and signed by a person, sits on the right side of that line. A hundred generic articles published in a month do not.

How long should an SEO article be?

As long as the question needs, and no longer. Google says it has no preferred word count, and padding an article to reach one makes it worse to read and harder to quote. Some questions are answered properly in a few hundred words; a definitive guide may need several thousand. Length is an outcome of the question, never a target in the brief.

How is this different from the SEO content system?

The SEO content system decides what to publish across a whole topic and runs the production machinery. SEO content writing is the standard and workflow for each article inside it. Many businesses start here and grow into the full system.

Will articles written this way show up in AI answers?

Answer-first sections, one question per heading, clear definitions and real examples are what make an article easy to quote, so the workflow builds them in. No one can guarantee a citation, any more than a ranking.

Let’s map what your content system could look like.

One conversation, an hour, no pitch deck. We look at what you already publish, where it stalls, and whether a system would actually help. If it wouldn’t, we’ll say so.

Start a Conversation