7 ChatGPT Prompts for Content Briefs by Article Type

Use one master ChatGPT prompt to define the brief's required fields, then attach a focused module for the article type or context you need. The master keeps output consistent. The modules change structure, evidence, competitor gaps, and voice without forcing your team to rebuild the workflow. Copy the seven templates below and replace each curly-brace variable.

What is this prompt library for?

This page is the working prompt library. The guide to AI content brief generation owns the end-to-end workflow, including research, intent, keyword strategy, outline design, quality control, tool selection, and scaling. Use that article to design the system. Return here when you need a prompt you can run.

A brief only earns its place when a writer can act on it without guessing. The BetterBriefs survey found that one-third of marketing spend may be wasted by poor briefs and misdirected work. A prompt cannot repair weak strategy, but it can force the strategist to make required decisions visible before drafting starts.

The seven templates use one master plus six modules. The sample output under the master is an example of the expected format, not another prompt.

How should you use the master prompt?

Fill every curly-brace variable before running the master. If you do not have a value, write NEEDS DATA and assign an owner. That instruction keeps missing inputs visible instead of inviting ChatGPT to invent search volumes, rankings, quotations, or source URLs.

Google's helpful content guidance asks whether content demonstrates first-hand expertise, serves an intended audience, and fulfills the reader's goal. Those questions belong in the brief. ChatGPT Search can search the web and cite sources, but browsing does not replace your approved keyword data, live SERP capture, customer evidence, or editorial judgment.

Prompt 1: Master content brief

You are a senior content strategist. Create a writer-ready content brief for the article defined below. Use the exact output fields requested. Do not invent statistics, search volumes, rankings, quotations, customer evidence, or source URLs. Write NEEDS DATA: {missing input and owner} wherever evidence is missing.

Use this context:

• Company and product: {one-sentence description}
• Intended audience: {role, company context, existing knowledge, and pain point}
• Primary keyword: {exact keyword}
• Supporting concepts: {terms and buyer questions}
• Search intent: {what the reader wants to accomplish}
• Article type: {how-to, comparison, thought leadership, or expert roundup}
• Target length: {range}
• Approved internal links: {URL plus relevance note for each}
• Desired reader action: {one action}
• SERP and competitor evidence: {pasted notes or NEEDS DATA}
• Brand and persona artifact: {pasted artifact or file name}

Return these fields:

• Three working titles, with the primary concept in at least two
• Meta title up to 60 characters
• Meta description from 150 to 160 characters
• One-sentence intent statement
• Audience summary covering role, pain point, and existing knowledge
• H2 and H3 outline with a one-sentence writer note under every heading
• Questions, examples, and evidence required under each heading
• Competitor coverage and gaps based only on supplied evidence
• Internal link placement for every approved URL
• External source types that need verification
• Reader action placement and wording direction
• Word-count range by major section
• A final quality-control list naming every unresolved input

Example output fragment

• Primary concept: sales call recording software
• Working title: Best Sales Call Recording Software for Five-Person SDR Teams
• Intent: Commercial investigation. The reader is shortlisting tools using transcription accuracy, CRM sync, and total cost.
• Audience: An SDR lead at a Series A SaaS company who knows the category and needs faster call coaching.
• Evidence gap: NEEDS DATA: strategist must provide current pricing and live competitor outlines.

Save the clean master. Add one article-type module and any evidence or brand module that applies. Do not rewrite the master for each assignment.

Which module should a how-to brief use?

The how-to module turns the outline into an executable sequence. It requires prerequisites, action-led headings, checkpoints, and troubleshooting because a generic list of steps leaves the writer to reconstruct the process.

Prompt 2: How-to article module

Apply this module to the completed master brief. Keep every master field, then replace the outline rules with the rules below.

• Reader goal: Complete {specific task} and verify that it worked.
• Prerequisites: List every tool, permission, account, file, and piece of prior knowledge required before step one.
• Step order: Follow the task's real chronological order. Do not reorder steps around keyword placement.
• Step headings: Start each step heading with an action verb.
• Step notes: For every step, state the action, expected result, and verification method.
• Evidence: Name the product documentation, screenshots, or original test needed to support each technical claim.
• Troubleshooting: Include at least three specific failure points with a diagnosis and correction.
• Next action: End with the logical follow-on task.

Return the revised outline plus a NEEDS DATA list for any prerequisite, screenshot, or verification result that the editor has not supplied.

Which module should a comparison brief use?

A comparison brief needs criteria before candidates. That sequence makes the verdict traceable and prevents unequal sections that quietly favor one option.

Prompt 3: Comparison article module

Apply this module to the completed master brief. Keep every master field, then replace the outline rules with the rules below.

• Reader decision: {purchase or commitment the reader must make}
• Shortlist: {options the reader is evaluating}
• Exclusions: {options already ruled out and why}
• Evaluation criteria: {criteria tied to the reader's constraints}
• Evidence standard: Use current first-party documentation for capabilities and pricing. Mark any unverified claim NEEDS DATA.
• Equal coverage: Give every option the same fields and comparable depth.
• Section pattern: Overview, strengths, limitations, verified pricing snapshot, and best-fit case.
• Verdict: Recommend an option by use case and state the tradeoff behind each recommendation.

Return the criteria before any option section. Flag any criterion that cannot be applied consistently across the full shortlist.

Which modules fit original and expert-led articles?

Thought leadership and expert roundups need different editorial controls. One starts with a defensible position. The other starts with a question and earns its value through synthesis.

Prompt 4: Thought-leadership article module

Apply this module to the completed master brief. Keep every master field, then replace the outline rules with the rules below.

• Audience belief to change: {specific conventional belief}
• Position: {what the article argues and why}
• Original evidence: {dataset, experiment, operating example, or named expert input}
• Strongest objection: {best opposing argument}
• Practical implication: {what the reader should do differently}

Build an outline that states the conventional view fairly, presents the evidence for the new position, answers the strongest objection, and ends with the operational implication. Do not mirror competitor outlines. If the position or original evidence is missing, stop and return NEEDS DATA instead of generating a generic explainer.

Prompt 5: Expert-roundup article module

Apply this module to the completed master brief. Keep every master field, then replace the outline rules with the rules below.

• Editorial question: {one question every contribution helps answer}
• Contributors: {name, role, company, and assigned angle}
• Response length: {word range per contributor}
• Outreach question: {exact question sent to each contributor}
• Attribution format: {one consistent name, role, and company pattern}

Build an outline with a short editorial frame, consistently formatted contributor sections, and a synthesis section. The synthesis must identify agreement, useful disagreement, and the editor's recommendation. Flag missing permissions, titles, quotations, and contributor context as NEEDS DATA.

How do you add competitor gaps without inventing a SERP?

Supply live evidence before asking for synthesis. This keeps the model focused on coverage and differentiation instead of producing a plausible-looking ranking page list.

Prompt 6: Competitor-gap module

Apply this module to the completed master brief. Analyze only the material supplied below.

• Primary query and market: {query, country, language, and device}
• Capture date: {date}
• Competitor one: {URL, H2 and H3 outline, notable claims, and evidence}
• Competitor two: {URL, H2 and H3 outline, notable claims, and evidence}
• Competitor three: {URL, H2 and H3 outline, notable claims, and evidence}
• Buyer questions: {People Also Ask, sales questions, support questions, or community questions}

Return:

• Table-stakes topics covered by all or most competitors
• Missing questions, examples, evidence, or decision criteria
• Repeated weak claims that need stronger sourcing
• One differentiated angle supported by the supplied evidence
• Outline changes required by that angle
• A NEEDS DATA list for any conclusion the evidence cannot support

Do not infer rankings, traffic, search volume, or freshness beyond the supplied capture.

How do you preserve brand voice and audience context?

Keep voice and persona in reusable context artifacts. A persistent artifact gives the team one version to maintain, while a specific AI brand voice guide gives the model observable rules it can follow and an editor can check.

Prompt 7: Brand and persona module

Apply this module to the completed master brief. Use the artifact below as a constraint on the brief, not as background to summarize.

• Audience role and company context: {role, team, company stage, and buying influence}
• Existing knowledge: {what the reader already understands}
• Active pain: {problem in the reader's own language}
• Decision stage: {awareness, consideration, or decision}
• Required vocabulary: {terms the audience uses}
• Avoided vocabulary: {terms, claims, and framing the brand rejects}
• Voice rules: {sentence style, point of view, tone, and evidence expectations}
• Approved examples: {two short excerpts that demonstrate the voice}

Return a brief that states what the writer can skip, what must be explained, which examples fit the reader, and which voice rules the draft must satisfy. Add a final voice check with pass or fail criteria for every rule. Do not describe the voice with generic labels such as professional or approachable.

OpenAI Projects can keep files and project instructions together across related chats. Store the canonical artifact in your own versioned system, then upload the approved version to the project so a platform setting never becomes the only copy.

How do you correct a generic brief?

Generic output usually points to a missing requirement, weak organization, or absent example. The PROMPT framework separates Persona, Requirements, Organization, Medium, Purpose, and Tone, which gives an editor a practical diagnostic.

Check the failed brief in this order:

• Confirm the intent and reader task are explicit.
• Replace broad fields with observable constraints and named evidence.
• Add the correct article-type module.
• Paste live competitor evidence when the brief makes a SERP claim.
• Mark missing information NEEDS DATA with an owner.
• Supply one approved example in the exact output format.

Examples often fix formatting and depth faster than more adjectives. Anthropic's few-shot prompting guidance recommends relevant, diverse examples and a consistent structure. The same method works here. Use a brief that writers completed without follow-up questions, then remove any customer facts that the new assignment should not inherit.

How can a team reuse these prompts safely?

Treat prompts as governed operating assets. The AI content workflow should define who owns the master, who can change modules, how examples are approved, and how unresolved inputs move back to the strategist.

ChatGPT Projects work well for a team or brand workspace because instructions and reference files remain together. If you need a controlled interface, OpenAI's GPT creation guidance explains how instructions, knowledge, and capabilities are configured. Its GPT FAQ also clarifies that GPT conversations begin fresh, so durable context must live in instructions and knowledge files.

Use a simple release process:

• Assign one owner for the master and each module.
• Keep a dated change log with the reason for every edit.
• Test changes against three representative assignments.
• Store approved examples separately from live customer data.
• Review NEEDS DATA patterns monthly to improve upstream inputs.

For one practical system each week, subscribe to the AI-Led Growth newsletter and get the templates, operating notes, and field lessons behind AI-native content workflows.