Content Writing · Long-form writing
The Best Prompt for an Article Outline
The best prompt for an article outline makes the model act like a commissioning editor: it fixes the one question the article answers, checks that your notes support a distinct angle, and then orders the sections to suit the type of article. Each section comes back with a working heading, its claim, the supporting material from your notes, the gaps and a word count. A research list at the end gathers everything you still need to find.
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.
- Works with
- ChatGPT, Claude, Gemini
- Difficulty
- Beginner
- Complexity
- Multi-part reply
- Variables
- 6
- 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.
6 variables · fields marked * matter most
# Role You are a commissioning editor preparing an outline that a writer could draft from without a further briefing. # Assignment - Topic: [Topic] - Article type: [Article type] - Reader: [Reader] - What the reader should take away: [What the reader should take away] - Target length: [Target length] # What I already have <notes> [Your notes and sources] </notes> These are my notes, facts and sources. They may be incomplete or out of order. # Task 1. Central question. Write the one question this article answers for the reader, and a one-sentence answer. Everything in the outline should serve that answer. 2. Angle check. Say in two sentences what a reader would get here that a generic article on the topic would not. If my notes do not support a distinct angle, say so and propose the closest one they do support. 3. Outline. Put the sections in the order that suits the article type: steps in sequence for a how-to, the strongest argument or the biggest objection first for an opinion piece, criteria before verdicts for a comparison. For each section give: - a working heading that states the section's point - the claim or job of the section in one sentence - the supporting material, quoting or pointing to my notes - anything missing, marked [RESEARCH: what to find] or [EXAMPLE: what kind] - an approximate word count 4. Opening and ending. Describe in a sentence each how the article should open and where it should stop. Do not write them. 5. Cuts. List anything in my notes you left out, and why. # Constraints - Do not supply facts, figures, quotes or sources that are not in my notes. Mark the gap instead. - Sections must not overlap. If two would make the same point, merge them. - No section exists only because articles usually have one. That includes background sections and closing summaries. - Word counts should add up to roughly the target length. - This is an outline. Do not draft paragraphs. # Output format Central question and answer, then the angle check, then the outline as a numbered list with the five fields under each heading, then the opening and ending notes, then cuts. Finish with a research list that gathers every [RESEARCH] and [EXAMPLE] marker in one place.
Use it when you are
- Organizing a pile of research notes into a structure before drafting
- Briefing a freelance writer or a colleague on an article you will not write yourself
- Finding out whether you have enough material for an article before committing time to it
- Restructuring a draft that wanders by outlining what it should have been
Why this prompt works
One question governs the whole outline
Outlines sprawl when the only organizing idea is the topic. Asking for a single central question with a one-sentence answer gives the model a test for every section: does this help answer it? Material that fails the test ends up in the cuts list, where you can see what was dropped and overrule the decision.
Section order depends on the kind of article
A how-to, an argument and a comparison are built differently, and a default heading sequence fits none of them well. Naming the ordering logic for each type stops the model from producing the same background, body, conclusion skeleton regardless of what you are writing.
Each section carries its own evidence
Requiring the supporting material for every section, quoted or referenced from your notes, shows at a glance which parts of the article are well supported and which rest on nothing. A section whose support field is mostly [RESEARCH] markers is a warning before you draft, not after.
It is told what not to write
The explicit ban on drafting paragraphs and on customary sections such as background and summary keeps the reply short enough to review in a minute. An outline you can scan and rearrange quickly is more useful than a half-written article you have to read in full.
How to use it
- Choose the article type first; it changes the order the sections come back in.
- Paste your notes as they are, with a source next to each fact where you have one.
- Check the central question and answer. If the answer is not what you want to argue, correct it and rerun before reading further.
- Scan the supporting material under each section and move, merge or delete sections that look thin.
- Take the research list away and fill the gaps, then add the findings to your notes and ask for a revised outline.
- Hand the final outline to the blog writing prompt, or draft from it yourself.
AI model compatibility
| Model | Status | Notes |
|---|---|---|
| ChatGPT | Expected to work | If it starts writing full paragraphs under each heading, remind it that the fields are one sentence each and ask it to reformat without adding content. |
| Claude | Expected to work | Notes are placed in XML-style tags, following Anthropic's advice to separate reference material from instructions. Long notes can go in as pasted documents inside the same tags. |
| Gemini | Expected to work | The numbered task list and the stated output format follow the explicit, ordered style Google's prompting guidance recommends; keep them intact when you edit. |
“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: A home-improvement writer planning an explainer from notes gathered from installers and a utility's published guidance. Illustrative inputs only.
# Role You are a commissioning editor preparing an outline that a writer could draft from without a further briefing. # Assignment - Topic: Choosing between a heat pump and a gas furnace - Article type: Comparison - Reader: Homeowners in cold climates whose furnace is near the end of its life - What the reader should take away: Know which three or four questions decide the choice for their house - Target length: 1,200 words # What I already have <notes> - Interviewed two local installers. Both said the house's insulation and ductwork matter more than the equipment brand. - Installer A: cold-climate heat pump models exist, but she always asks about backup heat. - Utility website has a rebate page (need to check current amounts). - My own house: kept the gas furnace as backup, heat pump does most of the work. - Need: running cost comparison. Do not have numbers yet. - Electrical panel capacity came up as a hidden cost in both interviews. </notes> These are my notes, facts and sources. They may be incomplete or out of order. # Task 1. Central question. Write the one question this article answers for the reader, and a one-sentence answer. Everything in the outline should serve that answer. 2. Angle check. Say in two sentences what a reader would get here that a generic article on the topic would not. If my notes do not support a distinct angle, say so and propose the closest one they do support. 3. Outline. Put the sections in the order that suits the article type: steps in sequence for a how-to, the strongest argument or the biggest objection first for an opinion piece, criteria before verdicts for a comparison. For each section give: - a working heading that states the section's point - the claim or job of the section in one sentence - the supporting material, quoting or pointing to my notes - anything missing, marked [RESEARCH: what to find] or [EXAMPLE: what kind] - an approximate word count 4. Opening and ending. Describe in a sentence each how the article should open and where it should stop. Do not write them. 5. Cuts. List anything in my notes you left out, and why. # Constraints - Do not supply facts, figures, quotes or sources that are not in my notes. Mark the gap instead. - Sections must not overlap. If two would make the same point, merge them. - No section exists only because articles usually have one. That includes background sections and closing summaries. - Word counts should add up to roughly the target length. - This is an outline. Do not draft paragraphs. # Output format Central question and answer, then the angle check, then the outline as a numbered list with the five fields under each heading, then the opening and ending notes, then cuts. Finish with a research list that gathers every [RESEARCH] and [EXAMPLE] marker in one place.
What a good response should contain
- A central question along the lines of which factors decide the choice for a particular house, not which technology is better in general
- Sections ordered criteria first (insulation and ducts, climate and backup heat, panel capacity, running costs), then how to weigh them
- Installer comments referenced as support under the relevant sections
- [RESEARCH] markers for running costs and current rebate amounts, with no figures filled in
- A cuts note or a placement decision for the writer's own-house anecdote
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 outline reflects your notes. It will not discover important subtopics you have not researched, though the angle check may point at them.
- Word counts per section are estimates for planning, not targets to hit exactly.
- It does not check your sources or whether the facts in your notes are correct.
- It is not built around search demand. For an article that must rank, start with an SEO content brief.
Alternative prompts
Content Writing
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.
- 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
Research
Research Paper and Report Summary Prompt
Summarizes a paper or report for a named reader: bottom line, method, key findings with exact figures, limits of the evidence, and what it means for them.
- ChatGPT
- Claude
- Gemini
Related resources
Technical guide
Prompt Chaining: Breaking a Task Into Linked Prompts
How to split a large task into a sequence of focused prompts, design the handoffs between them, and debug the chain when a step goes wrong.
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.
Frequently asked questions
Should I outline before or after doing the research?
Both. A first pass with thin notes shows you what the article needs and returns a research list. After you fill those gaps, a second pass produces the outline you will draft from. Running it twice is normal and usually quicker than researching without a structure.
How detailed should an outline be?
Detailed enough that someone else could write the article without asking you what each section is for. That means a point and its evidence per section, not full sentences of prose. If you find yourself writing paragraphs in the outline, you have started drafting.
Can I use this without any notes?
You can, but the result will be a generic structure with a [RESEARCH] marker under nearly every heading, because the prompt forbids the model from supplying its own facts. That is still a usable research plan, just not yet an outline of your article.
Sources and references
- Prompt engineering guide — OpenAI
- Creating helpful, reliable, people-first content — Google Search Central
Was this prompt useful?
No account needed. Do not include personal information.