SEO Automation

Keyword to Article Workflow

Turn one chosen keyword into a researched draft that's ready for a person to review.

SEO Automation EditorialPublished March 20, 2026Updated March 20, 20264 min read

Intermediate · Level 2–3 · Tools: DataForSEO, n8n, LLM, WordPress

  1. Keyword
  2. SERP data
  3. Research
  4. Outline
  5. Writing
  6. Fact checking
  7. Internal links
  8. Human review
  9. CMS

This workflow takes a single keyword you want to rank for and walks it all the way to a draft sitting in your CMS (the system you publish pages with, like WordPress). Software does the legwork. A person still checks the facts and decides whether it goes live.

Picture Priya, the only SEO person at a 40-person software company. She has a spreadsheet of 60 keywords everyone agrees are worth targeting, and she's written four articles this quarter. Each one eats most of a week: looking at who ranks, pulling notes together, outlining, writing, adding links, pasting it all into WordPress. Most of that is gathering and formatting, not thinking. This workflow hands the gathering and formatting to software, so her week goes into the parts only she can do.

At a glance

  • What it does: Turn one chosen keyword into a researched draft that's ready for a person to review.
  • Difficulty: Intermediate. Needs a few connected data sources and some care with prompts and review steps. Build a beginner workflow first if you can.
  • Automation level: Level 2–3 on the levels of SEO automation. In plain words: anything from a fixed chain of steps (Level 2) up to AI helping with judgement calls (Level 3), depending on how far you take it.
  • Steps: 9, from "Keyword" to "CMS".

What you'll need

Tools

  • DataForSEO: a paid service that sells keyword and search results data through an API (a way for one piece of software to ask another for data). See DataForSEO.
  • n8n: a workflow tool that connects your apps by dragging boxes around, so you don't need to write code. See n8n.
  • LLM: a large language model, the kind of AI behind ChatGPT, Claude and Gemini.
  • WordPress: the most widely used CMS. See the WordPress API guide.

What goes in

  • A keyword you've already chosen
  • Search results data for that keyword
  • Your brand voice and style notes
  • A list of your existing pages

What comes out

  • A research summary
  • An approved outline
  • A fact-checked draft
  • Suggested internal links
  • An unpublished CMS draft

How it works, step by step

  1. Keyword. You pick the keyword from your list. The workflow starts when you add it to a sheet or a form.
  2. SERP data. It pulls the current page of search results (the SERP) for that keyword: who ranks, their titles, and the questions people also ask.
  3. Research. An AI model reads the top results and summarises what they cover, what they miss, and what searchers seem to want.
  4. Outline. It proposes headings based on that research. You approve or tweak them before any writing happens.
  5. Writing. The AI drafts each section from the approved outline and your style notes.
  6. Fact checking. A second pass flags every claim, number and product detail that needs a source, so nothing slips through unchecked.
  7. Internal links. It suggests links to your existing pages that genuinely relate to the topic.
  8. Human review. An editor reads the whole thing, fixes the voice, confirms the facts and decides if it's good enough.
  9. CMS. The approved draft lands in WordPress as unpublished, with the title and meta description already filled in.

A worked example

Say the keyword is "how to reduce customer churn". The workflow pulls the top ten results and notices they all list the same generic tips, but none show how to spot churn early using product data. The outline it proposes leads with that gap.

The draft comes back at around 1,800 words with eleven claims flagged for checking. Two turn out to be invented statistics, and the editor cuts them. Four internal links are suggested and three are kept. In this example, Priya spends a couple of hours reviewing instead of most of a week writing, and the post sits in WordPress as a draft for her manager to approve.

Where AI helps, and where it doesn't

  • Summarising what the top-ranking pages cover, and what they skip.
  • Suggesting an outline built around what searchers actually ask.
  • Writing a first draft in your house style.
  • Spotting claims that need a source before an editor ever sees them.

AI shouldn't be the last word on facts, prices or anything about your own product. It states made-up details with total confidence, which is exactly why the fact-check step and the human review are there. It also shouldn't be the one pressing publish.

What to keep human

  • Someone checks every fact, number and quote against a real source.
  • An editor makes sure it sounds like your brand, not like generic AI writing.
  • A person decides whether it goes live.

For more on deciding how much to hand over, see the SEO autonomy dial and what not to automate.

What can go wrong

  • The AI invents facts or statistics that sound right but aren't.
  • The draft just mimics what already ranks and adds nothing new, so it ends up thin.
  • The new article targets the same keyword as an existing page, so the two compete with each other (known as cannibalization).

Start small

Skip the CMS step at first. Build just the first four steps: keyword in, search results pulled, research summary and outline out, delivered to a Google Doc. Write the article yourself from that outline. Once the outlines are reliably useful, add drafting, then linking, then the WordPress push.

Where to go next

Related pages