A blog post from keyword to published

What it is good at

• Fast at analysing what's currently ranking for a keyword and what those pages cover.
• Good at building a heading structure that answers a topic thoroughly.
• Can generate meta title and description options within character limits quickly.
• Useful for a final structure and readability check before publishing.

Where it falls short

• It can't tell you what will actually rank, search results depend on many factors beyond content quality.
• Content that only restates what's already ranking rarely outperforms the pages it copied.
• Real expertise and original examples still have to come from you, not the model.

How to use it for creative work

1. 01

   Research what's currently ranking

   Ask the model to summarise what the current top-ranking pages for your target keyword cover, and where they're thin, generic or outdated. This tells you what gap you're actually filling.

2. 02

   Decide your genuinely new angle

   Based on the gap, decide what you know or can show that those pages don't, a specific example, a different framework, first-hand data. This is the part no model can generate for you.

3. 03

   Build the heading structure

   Ask for a heading outline that covers the topic thoroughly based on common questions people ask about it, checked against what the top pages already cover so you're not just duplicating their structure.

4. 04

   Draft section by section

   Write or generate each section with your specific angle and examples included, rather than asking for the whole post in one shot, this keeps your real insight embedded rather than diluted.

5. 05

   Run the pre-publish check

   Before publishing, ask for a check on: keyword presence in the title and first paragraph, internal link opportunities, meta title and description length, and any sections that read as filler.

Ready-to-use prompts

For the search term '[keyword]', research what the current top-ranking pages cover. Summarise their common structure and flag anything they cover thinly, skip entirely, or get outdated on.

Build an H2/H3 heading outline for an article targeting '[keyword]' that answers the real questions people have about this topic, informed by this gap analysis: [paste]. Don't just copy competitor structure.

Write the section for the heading '[heading]' in an article about [topic]. Include this specific example/data point that's mine: [paste]. Tone: [describe]. Around [X] words.

Write five meta title options (under 60 characters) and three meta description options (under 155 characters) for an article about [topic], targeting the keyword '[keyword]', with a clear reason to click.

Review this draft against the keyword '[keyword]'. Check: is the keyword in the title and first paragraph, are there natural internal link opportunities, and is any section filler that could be cut. Draft: [paste]

Questions people ask

Will an AI-assisted blog post rank well?

Ranking depends on genuine usefulness, site authority and many other factors, AI can speed up research and drafting, but a post that just restates competitors rarely outranks them.

How much should I let AI write versus me?

Use it heavily for research, structure and drafting mechanics, but keep the genuinely new insight, examples and opinion as your own contribution.

Does keyword stuffing still matter?

No, natural keyword placement in the title, first paragraph and a few headings is enough; forcing the exact phrase repeatedly reads badly and doesn't help rankings the way it once did.