A Standard Operating Prompt (SOP) is the AI equivalent of a business process document: a tested, agreed-upon prompt that a team uses consistently for a specific type of task. When teams run AI without SOPs, every person gets slightly different outputs for the same task. When they run with SOPs, they get consistent, reviewable, improvable results.
What a Standard Operating Prompt Is
An SOP prompt is a prompt that has been:
- Written for a specific task type (not a general tool)
- Tested against real examples until it consistently produces acceptable output
- Agreed upon by the people who will use it
- Documented in a shared location with a version number
It's not a first draft or a starting point — it's a settled solution that's been validated.
The SOP Prompt Structure
A well-built SOP prompt has five components:
1. Persona
Tell the AI who it's being asked to act as. Not a fictional character, but a professional role: "You are an experienced technical writer preparing documentation for software developers." This sets the register, vocabulary level, and implicit standards for the output.
2. Context
What is the business context for this task? What does the AI need to know about your organisation, your audience, or your constraints to produce a relevant output? "Our documentation follows a problem/solution structure and uses British English. The target reader is a senior developer who values brevity over completeness."
3. Task
The specific instruction: what output is needed, from what input, in what form. "Transform the following meeting notes into a concise action item list. Each item should include: owner, action, deadline."
4. Format
Specify the exact output format. If you want a table, say table. If you want plain prose, say so. If you want a specific section structure, list it. SOPs should have consistent formats so that outputs from different team members can be used interchangeably.
5. Constraints and Quality Criteria
What should the output avoid? What's the definition of "good" for this specific task? "Do not include discussion points that were tabled without a decision. Flag any action item where the owner or deadline is unclear."
Three Example SOPs
SOP 1: Meeting Notes → Action Items
You are an executive assistant preparing a meeting summary.
Context: Internal team meeting notes, English-speaking team.
Task: Extract all action items from the following meeting notes.
Format: Numbered list. Each item: [OWNER] — [ACTION] — [DEADLINE].
Constraints: Only include items with a clear next action.
If owner or deadline is unclear, add [UNCLEAR] in that field.
[PASTE NOTES]SOP 2: Customer Request → Support Response
You are a customer support specialist for [COMPANY NAME].
Context: Our tone is helpful and direct. We do not use jargon.
We acknowledge problems without making promises we can't keep.
Task: Draft a response to the following customer message.
Format: Plain prose, 3–5 sentences. End with a clear next step.
Constraints: Do not promise a specific resolution timeline.
Do not use "unfortunately". Do not be defensive.
[PASTE CUSTOMER MESSAGE]SOP 3: Raw Data → Executive Summary
You are a business analyst preparing a one-page summary for a senior executive.
Context: The reader has 90 seconds. They want the key finding and one implication.
Task: Summarise the following data or report.
Format: Three sections — Key Finding (1 sentence), Context (2–3 sentences),
Implication for [FUNCTION: e.g., operations] (2–3 sentences).
Constraints: No jargon. Numbers rounded to one decimal place.
[PASTE DATA OR REPORT]Implementing SOPs with a Team
Start with the three tasks your team uses AI for most frequently. Build an SOP for each. Test it on ten real examples from the last month. Adjust based on failure patterns. Then document it in your team's knowledge base with a version date and an owner. Review every quarter — models change, task requirements change, and your SOP should reflect what you've learned.