Content Writing · Long-form writing
The Best Prompt for Writing a Blog Post
The best prompt for blog writing gives the model your point, your reader and your own source material, then makes it commit to a one-sentence thesis and a section plan before it drafts. This prompt also bans invented statistics and anecdotes, so gaps come back as [NEEDS EXAMPLE] and [VERIFY] markers. You get a plan, three title options, a full draft in markdown and a checklist of what to supply before publishing.
Blog Post Writing Prompt
Turns your own notes and experience into a focused blog post with one clear argument, and marks every spot where it needs a real example or a fact check.
- Works with
- ChatGPT, Claude, Gemini
- Difficulty
- Beginner
- Complexity
- Multi-part reply
- Variables
- 7
- Version
- 1.0
Customize this prompt
Fill in what you know. Anything left blank stays as a [bracketed placeholder]. Nothing you type leaves your browser.
7 variables · fields marked * matter most
# Role You are a blog writer and editor. You write for one specific reader, make one point per post, and would rather publish 600 useful words than 1,500 padded ones. # The post - Topic: [Topic] - Reader: [Reader] - The point I want to make: [The point you want to make] - What the reader should be able to do or decide afterwards: [What the reader should do or decide] - Voice: [Voice] - Maximum length: [Maximum length] # Source material <source_material> [Your notes, facts and examples] </source_material> This is what I know firsthand: experience, numbers, examples, quotes. It is the only place specific facts in the post may come from. # How to work 1. Thesis. State the post's argument in one sentence. If my point is too broad to argue in this length, narrow it and tell me how you narrowed it. 2. Plan. List the sections. For each, give the claim it makes and the source material that supports it. Drop any section that only repeats what this reader already knows. 3. Draft. Open with the reader's problem or with the answer itself, not with background. Start each section with its point, then support it. Where a concrete example would help and I have not supplied one, write [NEEDS EXAMPLE: what kind] instead of inventing one. 4. Close. Stop when the last useful thing has been said. A specific next step is fine; a summary of what the post just said is not. 5. Cut. Reread as the reader and remove every sentence they would skim past. # Constraints - Do not invent statistics, studies, quotes, customer stories or personal anecdotes. General knowledge is fine when stated plainly; add [VERIFY] after any claim I should check before publishing. - The length is a ceiling. Finish early if the argument is complete. - Subheadings say what the section argues, such as a claim or a question the reader has. Do not use labels such as Introduction or Conclusion. - Write in the voice above using plain words. No rhetorical-question openers, no scene-setting about how fast the world is changing, no stacked adjectives. - Use a list or table only where the content is a sequence or a comparison. Arguments stay in paragraphs. # Output format 1. Thesis and section plan, briefly. 2. Three title options that describe what the post delivers. 3. The post in markdown. 4. A checklist of every [NEEDS EXAMPLE] and [VERIFY] marker, with what I should supply for each.
Use it when you are
- Drafting a company or personal blog post from rough notes and a clear opinion
- Turning a talk, internal memo or customer conversation into a publishable article
- Writing an expertise-led post where the facts must come from you, not the model
- Producing a first draft that an editor or subject expert will finish
Why this prompt works
The thesis comes before the first paragraph
Most weak AI blog posts are a tour of a topic rather than an argument. Requiring a one-sentence thesis, and letting the model narrow a point that is too broad, gives every section something to serve. You also see the thesis first, so you can stop and correct the direction before reading a thousand words built on the wrong claim.
Your notes are the only source of specifics
A model with no material will produce plausible numbers and tidy anecdotes. Fencing your notes in tags and naming them as the only source of specific facts changes the job from inventing support to arranging yours. General knowledge is still allowed, but it has to carry a [VERIFY] flag where you should check it.
Gaps become markers, not fiction
The [NEEDS EXAMPLE: what kind] instruction gives the model something honest to do when the draft needs a story you did not provide. The closing checklist collects those markers, so finishing the post means answering a short list of requests rather than hunting for sentences that sound suspiciously convenient.
Length is a limit and the ending has a rule
Treating the word count as a ceiling removes the incentive to pad. The instruction to stop when the last useful thing is said, with no recap, targets the most common filler in generated posts: a closing section that repeats the subheadings back to the reader.
How to use it
- Write the point you want to make as a claim someone could disagree with. If you only have a subject, decide what you think about it first.
- Paste everything you know firsthand into the notes field: numbers, what happened, what you tried, quotes you may use. Do not tidy it.
- Pick a maximum length one size smaller than your instinct; the prompt treats it as a ceiling.
- Read the thesis and section plan before the draft. If the thesis is wrong, correct it in one line and ask for a new plan.
- Work through the closing checklist: supply each missing example, check each [VERIFY] claim, then ask the model to revise only those passages.
- Read the final draft aloud once and rewrite any sentence that does not sound like you.
AI model compatibility
| Model | Status | Notes |
|---|---|---|
| ChatGPT | Expected to work | If the draft drifts from the plan it wrote in step 2, quote the plan back and ask for the affected section to be rewritten against it. |
| Claude | Expected to work | The source material is already wrapped in XML-style tags, which Anthropic's guidance recommends for separating supplied documents from instructions. Keep the tags when you paste long notes. |
| Gemini | Expected to work | Leave the numbered steps in order. Google's prompting guidance favors explicit sequential instructions and a stated output format. |
“Expected to work” means the prompt uses plain, model-independent structure and has been reviewed for that — not that it has been run on that model. A model is marked Tested only when a test record is published below.
Worked example
Scenario: An engineering manager writing for a company blog about what they changed in remote onboarding. Illustrative inputs, not a recorded output.
# Role You are a blog writer and editor. You write for one specific reader, make one point per post, and would rather publish 600 useful words than 1,500 padded ones. # The post - Topic: Onboarding new engineers remotely - Reader: Engineering managers at startups with 10 to 50 developers who have hired remotely at least once - The point I want to make: A written first-week plan with one small shipped change matters more than assigning a buddy - What the reader should be able to do or decide afterwards: Write a first-week plan before their next hire starts - Voice: Conversational and direct - Maximum length: 900 words # Source material <source_material> - We used to assign a buddy and nothing else. New hires told us in retros they did not know what they were supposed to be doing in week one. - Now every hire gets a one-page plan: day 1 environment setup with a named helper, day 2-3 a pre-picked small bug, day 5 demo at team meeting. - We keep a list of starter bugs tagged in the tracker. - Buddy still exists but their job is defined: 20 minutes a day, first week only. - One hire's first change was fixing a typo in an error message customers saw. They still bring it up. </source_material> This is what I know firsthand: experience, numbers, examples, quotes. It is the only place specific facts in the post may come from. # How to work 1. Thesis. State the post's argument in one sentence. If my point is too broad to argue in this length, narrow it and tell me how you narrowed it. 2. Plan. List the sections. For each, give the claim it makes and the source material that supports it. Drop any section that only repeats what this reader already knows. 3. Draft. Open with the reader's problem or with the answer itself, not with background. Start each section with its point, then support it. Where a concrete example would help and I have not supplied one, write [NEEDS EXAMPLE: what kind] instead of inventing one. 4. Close. Stop when the last useful thing has been said. A specific next step is fine; a summary of what the post just said is not. 5. Cut. Reread as the reader and remove every sentence they would skim past. # Constraints - Do not invent statistics, studies, quotes, customer stories or personal anecdotes. General knowledge is fine when stated plainly; add [VERIFY] after any claim I should check before publishing. - The length is a ceiling. Finish early if the argument is complete. - Subheadings say what the section argues, such as a claim or a question the reader has. Do not use labels such as Introduction or Conclusion. - Write in the voice above using plain words. No rhetorical-question openers, no scene-setting about how fast the world is changing, no stacked adjectives. - Use a list or table only where the content is a sequence or a comparison. Arguments stay in paragraphs. # Output format 1. Thesis and section plan, briefly. 2. Three title options that describe what the post delivers. 3. The post in markdown. 4. A checklist of every [NEEDS EXAMPLE] and [VERIFY] marker, with what I should supply for each.
What a good response should contain
- A one-sentence thesis close to the supplied point, possibly narrowed to the first week only
- A section plan where each claim is tied to a specific line from the notes
- An opening that starts with the new hire who does not know what to do, not with remarks about remote work in general
- No statistics about onboarding or retention, since none were supplied, and [VERIFY] on any general claim it adds
- A checklist asking for things such as a copy of the one-page plan or a second concrete starter bug
This is an illustrative scenario showing how to fill in the prompt and what to look for. It is not a recorded model output. Recorded outputs will be published here once this prompt has a test record.
Testing and editorial review
Status: Untested. Reviewed for structure and clarity by the editorial team. No recorded model test yet.
No test record has been published for this prompt yet. When it is tested, this section will show the model and version, the date, the inputs, the actual output and notes against our seven-point rubric. We do not publish numerical scores.
| Author | TheBestPromptFor Editorial Team — The editorial team writes, reviews and maintains the prompts and guides on TheBestPromptFor. Prompts are drafted with AI assistance and reviewed against published editorial standards. |
|---|---|
| Published | November 15, 2026 |
| Last reviewed | November 15, 2026 |
| Last tested | Not yet tested |
| Version history |
|
Limitations and things to check
- The post is only as specific as your notes. Thin input produces a short, general draft with many markers.
- The model cannot confirm that your facts are accurate or that you have permission to publish quotes and customer details.
- Voice options are broad. Expect to edit for your own phrasing, or paste a sample of your writing and ask it to match that.
- It does not do keyword research or search optimization. Use an SEO content brief first if the post has to rank for a query.
Alternative prompts
Content Writing
Article Outline Prompt
Builds a section-by-section outline around one central question, with the evidence for each section and a list of what still needs research.
- ChatGPT
- Claude
- Gemini
Content Writing
Editing and Proofreading Prompt
Edits a draft at the level you choose, protects the author's voice, and returns a change log plus queries instead of silent rewrites.
- ChatGPT
- Claude
- Gemini
Content Writing
Headline Writing Prompt
Generates headlines from several distinct angles, checks each against accuracy and length tests, and recommends three with reasons.
- ChatGPT
- Claude
- Gemini
SEO
SEO Content Brief Prompt
Produces a writer-ready brief: reader intent, the angle only you can offer, a purposeful outline, evidence to gather and clear acceptance criteria.
- ChatGPT
- Claude
- Gemini
- Perplexity
Related resources
Guide
How to Write a Good AI Prompt
The five parts of a complete prompt (goal, context, task, constraints, output format), with a worked rewrite and a pre-send checklist.
Guide
How to Improve AI Responses
A diagnostic approach to weak AI answers: name the fault, trace it to the gap in your prompt, then follow up precisely or rewrite and start again.
Frequently asked questions
What if I do not have notes, only a topic?
Spend ten minutes writing down what you have seen, done or measured on the topic, even as fragments. Without that, the prompt will still run, but the draft will lean on general knowledge and return a long list of [NEEDS EXAMPLE] markers. That list is a fair picture of what the post is missing.
Can I use this for SEO articles?
Yes, as the drafting step. Decide the search query, the intent and the required subtopics with a content brief first, then paste the brief into the notes field. This prompt focuses on argument and accuracy and does not choose keywords or headings for search.
Why does it ask for a point instead of just a topic?
A topic can be covered forever; a point can be made and finished. Giving the model a claim to argue lets it decide what belongs in the post and what does not, which is what keeps the draft short and gives the reader a reason to keep going.
How should I handle the [VERIFY] markers?
Treat each one as a claim that came from the model's general knowledge rather than from you. Check it against a source you trust, then either keep it, correct it or cut it. Remove every marker before publishing; none should survive into the final post.
Sources and references
- Prompt engineering overview — Anthropic
- Creating helpful, reliable, people-first content — Google Search Central
Was this prompt useful?
No account needed. Do not include personal information.