Keep a running SOP library from how you already work
Fewer 'how do we do X?' interruptions
Illustrative example — a composite of common setups, not a specific business's story.
- Difficulty
- No-code
- Setup time
- 20 minutes per SOP
- Monthly cost
- $0-$25
- Impact
- Medium
The short version
Record a quick screen or voice walkthrough; an assistant turns it into a step-by-step SOP stored in a shared folder.
What they did
Record a quick screen or voice walkthrough; an assistant turns it into a step-by-step SOP stored in a shared folder. The owner started with a small pilot: one workflow, a few days of supervised use where every output was reviewed before it went out, then gradually let routine cases run with lighter oversight.
Agent/tool used: ChatGPT / Scribe
How it went
In a typical rollout like this: fewer 'how do we do x?' interruptions. Most of the value comes from consistency, the task happens every time instead of when someone remembers. Expect a week or two of tuning the prompt and rules.
Worth knowing
AI drafts can be wrong or off-tone. Keep a human approving anything customer-facing at first, don't paste sensitive data (card numbers, health records, SSNs) into general chat tools, and check each tool's data and privacy terms.
Copy-paste prompt
Turn this walkthrough transcript into a standard operating procedure: [paste]. Include purpose, when to use it, tools needed, numbered steps, common mistakes, and who to ask. Plain language for a new hire.
Replace anything in [brackets] with your own details.
Step-by-step setup
- Write down how you handle this today, in 5-10 bullets.
- Connect or export data from Google Workspace, Slack.
- Set up ChatGPT / Scribe and paste in the prompt, filling in the [placeholders].
- Test it on 5-10 real past examples and fix anything that's off.
- Run it with you approving every output for one to two weeks.
- Once you trust it, automate the routine cases and spot-check weekly.
Published 2026-09-20. Want help setting this up?