ChatGPT Standard Operating Procedure (SOP) Writer Prompt
You are an operations specialist who writes SOPs clear enough that any qualified team member can follow without addition
Category
💼 Business
Difficulty
Intermediate
Models
3
Last Updated
2026-06-28
Works with
📄 Example output
⚠️ Common Mistakes
❓ FAQ
⚙️ Fill in your variables
📋 Prompt
You are an operations specialist who writes SOPs clear enough that any qualified team member can follow without additional training.
Process: [process name]
Performer: [who performs it — role, not name]
Frequency: [frequency — daily/weekly/per event]
Tools: [tools required]
Task: Write a complete Standard Operating Procedure:
HEADER: Title, SOP number, version, date, owner, next review date
PURPOSE: Why this process exists (1–2 sentences)
SCOPE: Who follows this and when
PRE-REQUISITES: What must be true before starting
STEP-BY-STEP INSTRUCTIONS: Numbered, one action per step, decision points as 'if X then Y, else Z', tools in [brackets]
QUALITY CHECKS: How to verify the process was done correctly
EXCEPTION HANDLING: What to do when it goes wrong
COMPLETION CRITERIA: How you know you're done
REVISION HISTORY: Table for tracking changes
Process: [process name]
Performer: [who performs it — role, not name]
Frequency: [frequency — daily/weekly/per event]
Tools: [tools required]
Task: Write a complete Standard Operating Procedure:
HEADER: Title, SOP number, version, date, owner, next review date
PURPOSE: Why this process exists (1–2 sentences)
SCOPE: Who follows this and when
PRE-REQUISITES: What must be true before starting
STEP-BY-STEP INSTRUCTIONS: Numbered, one action per step, decision points as 'if X then Y, else Z', tools in [brackets]
QUALITY CHECKS: How to verify the process was done correctly
EXCEPTION HANDLING: What to do when it goes wrong
COMPLETION CRITERIA: How you know you're done
REVISION HISTORY: Table for tracking changes
SOP-001: AI Prompt Library Weekly Quality Review
Version 1.0 | Date: 2026-06-28 | Owner: Content Lead | Review: 2026-12-28
PURPOSE: Ensure all new AI prompts meet quality standards before being published to the live library.
SCOPE: Performed by the Content Lead every Monday, covering all prompts submitted the previous week.
PRE-REQUISITES:
• GitHub access to farhan707/toolsnova repository
• Access to ToolsNova staging environment
• Quality checklist (Appendix A)
STEP-BY-STEP:
1. Open [GitHub] → farhan707/toolsnova → data/prompts.json
2. Filter for entries with lastUpdated = previous week
3. For each prompt:
3a. Verify all required fields present: slug, title, category, models, variables, prompt, example_output, tips (min 4), mistakes (min 4), faq (min 4)
3b. If any field missing → flag in [Notion] → proceed to next prompt
3c. Test prompt in [ChatGPT] using example variable values
3d. If output quality meets standard → mark 'Approved'
3e. If output quality is poor → flag for revision with specific note
4. For all flagged prompts → create revision tasks in [Notion]
5. Update prompts.json with all approved status changes
6. Run: python3 generator.py in [Terminal]
7. Push to staging: git push staging
8. Verify 3 random prompt pages load correctly
9. Merge to main: git push origin main → triggers Cloudflare auto-deploy
QUALITY CHECKS:
✅ All required fields present in every approved prompt
✅ Example output is genuine AI output, not manually written
✅ All pages load correctly on staging
COMPLETION: Process complete when all prompts reviewed + revisions logged + approved prompts live on production
Version 1.0 | Date: 2026-06-28 | Owner: Content Lead | Review: 2026-12-28
PURPOSE: Ensure all new AI prompts meet quality standards before being published to the live library.
SCOPE: Performed by the Content Lead every Monday, covering all prompts submitted the previous week.
PRE-REQUISITES:
• GitHub access to farhan707/toolsnova repository
• Access to ToolsNova staging environment
• Quality checklist (Appendix A)
STEP-BY-STEP:
1. Open [GitHub] → farhan707/toolsnova → data/prompts.json
2. Filter for entries with lastUpdated = previous week
3. For each prompt:
3a. Verify all required fields present: slug, title, category, models, variables, prompt, example_output, tips (min 4), mistakes (min 4), faq (min 4)
3b. If any field missing → flag in [Notion] → proceed to next prompt
3c. Test prompt in [ChatGPT] using example variable values
3d. If output quality meets standard → mark 'Approved'
3e. If output quality is poor → flag for revision with specific note
4. For all flagged prompts → create revision tasks in [Notion]
5. Update prompts.json with all approved status changes
6. Run: python3 generator.py in [Terminal]
7. Push to staging: git push staging
8. Verify 3 random prompt pages load correctly
9. Merge to main: git push origin main → triggers Cloudflare auto-deploy
QUALITY CHECKS:
✅ All required fields present in every approved prompt
✅ Example output is genuine AI output, not manually written
✅ All pages load correctly on staging
COMPLETION: Process complete when all prompts reviewed + revisions logged + approved prompts live on production
🏆
💡 Pro Tips
Best model for this prompt
Claude
Claude (Opus 4 / Sonnet 4)
Write for the least-experienced qualified person who will ever follow it — if they can follow without asking questions, it's done
Test every SOP by having someone else follow it verbatim — gaps reveal themselves immediately
Screenshots are worth 1,000 words for digital SOPs — show UI actions, don't just describe them
Build the review date into the SOP header and actually set a calendar reminder for it
Multi-action steps ('click X, then go to Y, then do Z') — one action per step prevents getting lost
Assumed knowledge ('configure the settings') — specify which settings and exact values
No exception handling — real processes go wrong; SOPs without troubleshooting sections get abandoned in practice
Written by the expert without being tested by the novice — experts skip steps they consider obvious
- How detailed should an SOP be?Detailed enough that your target user can follow it without asking questions. For experienced teams: higher-level. For new hires or complex processes: fully granular. Right test: have someone follow it cold, note where they get stuck.
- What format should SOPs be in?Notion or Confluence for searchable, linkable digital SOPs. Word/Google Docs for formal documentation. Video walkthroughs for complex visual processes. Use whatever your team will actually read and maintain.
- How often should SOPs be reviewed?Annual minimum. After any process change: immediate update. After a process failure: root cause analysis and SOP update. Build the review date into the header so it's never forgotten.
- Is there a difference between an SOP and a checklist?SOPs describe the complete process with context, decision points, and exception handling. Checklists are the simplified task-by-task execution tool derived from the SOP. Most processes benefit from having both: the SOP for training, the checklist for daily execution.