Prompts for Creating SOPs and Process Documentation

A Standard Operating Procedure is a recipe for doing the same task the same way every time. Most teams never write theirs down because it feels like a chore. AI removes that excuse — if you can describe a process out loud, AI can turn it into a clean, repeatable SOP in under fifteen minutes.

1. Introduction

SOPs are the connective tissue of any team that has grown past five people. They let you onboard new hires in days instead of months, they catch mistakes before they happen, and they survive the inevitable day when your most experienced operator goes on leave. Yet most SOPs live in someone's head. The reason is simple: writing them is boring and time-consuming.

AI changes the economics. You describe the process in your own words — messy, half-remembered, with all the edge cases — and the AI produces a structured, numbered, role-aware document. This tutorial shows you how to prompt for SOPs that are actually used, not just filed.

2. The Concept Explained

A useful SOP answers five questions: who runs this process, when does it run, what are the steps in order, what could go wrong at each step, and how do we know it worked. Almost every weak SOP fails one of those five.

Think of an SOP as a flight checklist. Pilots do not memorise every action — they read the checklist out loud, in order, every single flight, regardless of experience. The checklist exists because human memory is unreliable under pressure. Your team's customer-onboarding process or invoice-approval workflow needs the same protection.

3. The Problem Without This Technique

write an SOP for our customer onboarding process

4. The Solution

Act as a senior operations manager writing internal
documentation for a SaaS company called Brightline HR
(payroll software for SMBs in India).

Document type: Standard Operating Procedure (SOP).

Process: Onboarding a new paid customer from contract
signed to first payroll run.

Roles involved: account executive (signs the deal),
implementation specialist (configures the account),
customer success manager (handover after first run).

I am going to describe the process in messy first-person
notes below. Your job is to turn it into a clean SOP
with this structure:

1. Purpose (2 sentences).
2. When this SOP applies (trigger and scope).
3. Roles and responsibilities (table).
4. Step-by-step procedure (numbered list — each step
   includes the role responsible, the action, and the
   expected output).
5. Common edge cases and how to handle them (at least
   three).
6. Success criteria — how we know onboarding is complete.
7. Document owner and review cadence.

Rules:
- Keep steps small enough that a new hire could execute
  them without asking questions.
- Use plain language, no internal acronyms.
- If my notes are missing information for a step, write
  "[needs clarification — owner: ops manager]" instead
  of guessing.

My notes:
"""
... paste messy notes here ...
"""

5. Step-by-Step Breakdown

1. Pick one process at a time. SOPs work because they are narrow. "How we close monthly books" is an SOP. "How finance operates" is not.
2. Brain-dump the process out loud first. Voice-to-text or rough bullet points. Do not try to clean it up — that is the AI's job.
3. Name every role. An SOP without role names is a wish list. Each step needs an owner.
4. Demand the seven-section structure. Purpose, applies-when, roles, steps, edge cases, success criteria, owner. Lock the shape.
5. Force gap-flagging. "If my notes are missing information, write [needs clarification]". This stops AI from inventing steps that do not exist in your company.
6. Pilot before publishing. Have one person follow the SOP end-to-end. Every question they have is a gap to close in the next version.

6. Practice Exercises

7. Key Takeaways