ChatGPT Prompts

Example

Replace {TONE} with “direct but respectful,” {RECIPIENT} with “my manager,” {TOPIC} with “requesting a deadline extension on the Q3 report,” {LENGTH} with “under 150 words,” and {DESIRED OUTCOME} with “get approval for a 5 day extension without sounding unreliable.”

Why it Works

Specifying tone, length, and the desired outcome eliminates the three most common problems with AI drafted emails: wrong tone, too long, and unclear purpose. The “no filler phrases” constraint forces ChatGPT to start with substance.

Output Format

Subject line followed by email body.

Example

Replace {MEETING TYPE} with “quarterly budget review,” {TIME FRAME} with “2 hours,” {ATTENDEES} with “the CFO and department heads,” and {AGENDA ITEMS} with “headcount requests for Q4, vendor contract renewals, and the marketing automation budget.”

Why it Works

The structured output (3 points, 2 objections, 1 question) forces ChatGPT to prioritize rather than dump everything it knows. The word limit prevents the brief from becoming longer than the meeting itself.

Output Format

Numbered sections with headers.

Example

Replace {TOPIC} with “remote work productivity statistics,” and {TIME RANGE} with “the last 6 months.”

Why it Works

The table format makes findings scannable. Requiring sources and dates forces ChatGPT to use its web search capability rather than generating claims from training data. The source type filter reduces low quality results.

Output Format

HTML table with 4 columns.

Example

Replace the options with “Asana, Monday, and ClickUp” for “our team’s project management tool.” Priorities: “ease of onboarding for non technical users, integration with Google Workspace, and price for a team of 25.”

Why it Works

Forcing ChatGPT to show the weighted math makes the recommendation transparent and auditable. You can adjust scores or weights and rerun the analysis instantly. This turns a subjective decision into a structured evaluation.

Output Format

Weighted scoring table followed by a recommendation paragraph.

Example

Replace {CONTENT TYPE} with “blog post,” {NUMBER} with “5,” {OUTPUT FORMAT} with “LinkedIn posts,” {PLATFORM} with “LinkedIn,” {LENGTH} with “under 200 words each,” and {GOAL} with “engagement and comments from product managers.”

Why it Works

The “do not repeat” constraint forces ChatGPT to extract different angles from your source material instead of rewriting the same summary five times. Specifying the platform and audience means each piece uses the right conventions.

Output Format

Numbered posts, each with a hook line and body.

Example

Replace {LANGUAGE} with “Python,” {INTENDED BEHAVIOR} with “filter a list of dictionaries by a date range,” and {ACTUAL BEHAVIOR} with “returns an empty list even when matching records exist.”

Why it Works

Describing both intended and actual behavior gives ChatGPT the full debugging context. The structured output (one sentence explanation, corrected code, prevention tip) stops it from writing a 500 word essay when you just need the fix.

Output Format

Bug explanation, corrected code block, prevention tip.

Example

Replace {DECISION} with “switching from Slack to Microsoft Teams” and {CONTEXT} with “We are a 40 person marketing agency. Half the team uses Macs. We already pay for Google Workspace but not Microsoft 365.”

Why it Works

The “do not soften” instruction overrides ChatGPT’s tendency to present both sides as roughly equal. The “so what” requirement forces practical relevance instead of abstract points. The single factor recommendation cuts through analysis paralysis.

Output Format

Two numbered lists followed by a recommendation sentence.

Example

Upload a PDF and replace {DOCUMENT TYPE} with “vendor proposal for our CRM migration” and {LENGTH} with “42 pages.”

Why it Works

The three layer structure lets different audiences use the summary differently. The C suite reads layer one. Your manager reads layer two. You reference layer three for details. The contradiction flag catches issues that a straightforward summary would miss.

Output Format

Three clearly labeled sections with increasing detail.

Example

Replace {CONTENT TYPE} with “blog post,” {TOPIC} with “why annual performance reviews are outdated,” {AUDIENCE} with “HR managers at mid size companies,” {WORD COUNT} with “1,200 words,” {TONE} with “authoritative but not academic,” {STRUCTURE} with “problem, evidence, alternatives, implementation steps,” and {NUMBER} with “4.”

Why it Works

The banned phrases list eliminates the most common AI writing cliches. Requiring specific examples forces research rather than generalities. Defining the structure prevents the aimless “here are 7 reasons” format that plagues AI content.

Output Format

Formatted article with headings.

Example

Upload a CSV and replace {DESCRIPTION OF DATA} with “monthly sales by region for the last 12 months, columns are Date, Region, Revenue, Units Sold, and Sales Rep.”

Why it Works

Describing the data structure helps ChatGPT parse the file correctly. The three part output (trends, anomalies, actions) mirrors how a data analyst would present findings to a stakeholder. Asking for actionable recommendations prevents the analysis from ending with “further investigation is needed.”

Output Format

Tables for trends and anomalies, numbered list for recommendations.

Example

Replace {JOB TITLE} with “Senior Product Manager,” {COMPANY TYPE} with “a Series B fintech startup with 150 employees,” and {BRIEF BACKGROUND} with “6 years in product at enterprise SaaS companies, led a team of 4 PMs, shipped payment integrations and reporting dashboards.”

Why it Works

The “why they are asking it” layer reveals the subtext behind interview questions, which changes how you answer. The framework gives structure without scripting. The personalized example makes preparation concrete rather than theoretical.

Output Format

8 numbered blocks, each with question, intent, framework, and example.

Example

Replace {LIST OF TASKS} with your actual task list, {HOURS} with “9am to 6pm with a 1 hour lunch,” {TIME OF DAY} with “morning,” and {HOURS} of focus time with “2 hours.”

Why it Works

The “flag anything unrealistic” instruction turns ChatGPT from a passive scheduler into an active workload advisor. It will tell you when you have committed to more than your hours allow, which is the most valuable output a planner can give.

Output Format

Daily table with time blocks, followed by a feasibility note.