Ten prompt patterns for everyday knowledge work

Many workplace prompts can start from a small set of reusable patterns. The pattern gives the request a shape; you still customise it for the task, the evidence available, and the consequences of getting it wrong.

These ten patterns are useful starting points across writing, analysis, decision support, learning, and coding. They are not an exhaustive taxonomy, and the same wording will not work equally well across every model or task.

A pattern can make intent easier to preserve than a copied prompt whose assumptions you do not know. Choose the pattern for the job, then add context, constraints, and validation for the specific task.

1. The interview-me pattern

When you want the model to help you with something but you have not fully articulated what you want, let it interview you.

Before drafting anything, ask me [N] questions you would need answered to do this well. Wait for my answers before continuing.

This pattern is useful when important context is missing or you have not yet defined the desired outcome. The questions can help you sharpen the request, but review them: the model may ask irrelevant questions or miss a constraint that matters.

Use it for: drafting any complex piece, planning a project, getting unstuck on a decision, designing anything.

2. The critique-and-revise pattern

When you have a draft (yours or AI-generated), do not ask for “a better version.” Ask for critique, then ask for the specific revision.

Read this draft carefully. Tell me:

1. What is working — the parts I should not change.
2. What is weak — specific lines or paragraphs, with quoted text.
3. What is missing.
4. Anything that contradicts itself or sounds wrong.

Do not rewrite anything yet.

After the critique, you can either revise yourself or ask the model to do specific revisions (“rewrite the second paragraph based on your critique, keeping the others”).

Use it for: improving any piece of writing, code review, analysis review, slide review.

3. The three-variants pattern

For anything where the right answer is judgement-dependent, generate three options and pick.

Produce three versions:

• Version A: [specific constraint, e.g., short and direct, 50 words]
• Version B: [specific constraint, e.g., warm and explanatory, 150 words]
• Version C: [specific constraint, e.g., formal and detailed, 250 words]

Label each clearly.

Comparing constrained options can make trade-offs visible. It also costs more review time, so use it when several framings are genuinely useful rather than as an automatic rule.

Use it for: emails, social posts, headlines, taglines, slide titles, code approaches, and other tasks with multiple valid framings. Do not use generated variants to rank people or make consequential eligibility decisions.

4. The devil’s-advocate pattern

Some assistants have been measured preferring responses that match a user’s stated view, even when that view is wrong (Sharma et al., 2023). A counterargument prompt can expose another line of reasoning, although it does not make either side true:

[After getting a response.] Now play devil’s advocate. Take your own answer and produce the strongest credible counter-argument. What would the smartest critic of this position say?

This can surface weak assumptions or unsupported optimism in the earlier response. Verify the counterargument too; a model can invent a persuasive objection as readily as a persuasive defence.

Use it for: decisions, strategy, business cases, anything where you are about to act on what the model said.

Do not use it as theatre. If the decision is factual or regulated, the counterargument is not a substitute for checking sources, policy, or legal requirements.

5. The structured-template pattern

When you want a specific output shape, do not describe the shape — show it.

Respond in exactly this format:

Decision: [your recommendation]

Top three reasons for:

1. […]
2. […]
3. […]

Top three reasons against:

1. […]
2. […]
3. […]

What would change my answer: […]

Confidence level: [Low / Medium / High] because […]

The template gives the model a target schema and makes omissions easier to notice. It does not guarantee valid or consistent output, so validate required fields and allowed values when another system will consume the result.

Use it for: analytical decisions, structured reports, batch processing where you need consistent format, anything you want to paste into a doc and have look uniform.

6. The step-back pattern

When you are about to ask a specific question, first ask the broader one.

Before answering my specific question, take a step back and tell me: what is the general principle or framework that applies here? Then apply it to my specific case.

On an analytical question, this can separate the proposed framework from its application to your case. Check both: a neat framework can still be irrelevant, incomplete, or based on unsupported claims.

Use it for: technical questions, strategic decisions, learning, anything where the right answer depends on a general principle being applied correctly to a specific situation.

7. The persona pattern

Assigning a role can focus the model’s behavior and tone. Anthropic’s prompting best practices recommend role prompting for that purpose, but a persona does not give the model real qualifications or verified expertise.

Act as [a specific kind of collaborator]. Communicate in [a specific style]. Follow [specific habits and verification rules].

A few starting examples:

• “Explain Estonian tax terminology in plain language. Cite official Tax and Customs Board guidance, and flag anything a qualified accountant should review.”
• “You are a careful copy editor who only changes what genuinely needs changing.”
• “Act as a skeptical investment sparring partner. Ask hard questions and separate evidence from assumptions.”
• “Act as a patient tutor. Check my understanding before moving to the next topic.”

Be specific about the task, communication style, constraints, and validation. Invented credentials such as “fifteen years of incident response” may make an answer sound authoritative without making it more reliable.

For tax, legal, medical, security, or regulatory work, verify the answer against authoritative sources or a qualified professional. Persona prompting controls presentation and behavior; it is not evidence that the content is correct.

Use it for: anything where the kind of response matters as much as the content.

8. The role-play pattern

A variation on persona: have the model play a person rather than just an expert, and have a conversation with them.

Role-play as my prospective client. They are a marketing director at a mid-size B2B company. They are friendly but skeptical and have been burned before. I am about to pitch them on hiring my agency. Ask me questions and react the way they would. Stay in character. Push back if my pitch sounds vague or oversold.

Use it for: interview prep, sales practice, hard-conversation rehearsal, and exploring possible customer reactions. Treat the dialogue as generated practice, not evidence of how a real person or group will respond.

9. The contrast-pair pattern

When you cannot quite describe what you want, show what you do and do not want.

Here is an example of writing that hits the tone I want: [paste good example]

Here is an example of writing that misses the tone: [paste bad example]

The differences I care about are [your observation]. Now write [your task] in the style of the first one.

The contrast makes the dimension you care about more concrete than adjectives alone. Check the result for unwanted copying, confidential material, and style features you did not intend to preserve.

Use it for: brand voice, code style, design taste, anything where the right answer is “more like this, less like that.”

10. The output-template pattern

A formal version of structured-template (#5) for repeatable batch work. Define a strict template, then apply it across many items.

For each item I give you, produce a response in this exact JSON format:

{
  "summary": "...",
  "priority": "...",
  "category": "...",
  "next_action": "..."
}

Priority must be one of: “high”, “medium”, “low”. Category must be one of: “billing”, “technical”, “feature_request”, “other”. Wait for each item before responding.

Strict templates are useful in structured workflows such as classification, extraction, and integrations. Specificity can reduce ambiguity, but reliability comes from schema validation, error handling, and measured performance on representative inputs.

Use it for: classification, extraction, batch processing, anything where downstream tools need consistent format.

For automation, pair this pattern with validation. If the output is not valid JSON, contains a category outside the allowed list, or leaves a required field blank, the workflow should stop or route to human review.

Putting them together: a real workflow

The patterns are not exclusive. A typical complex task chains several together.

Suppose you are writing a proposal for a new company policy. The workflow might look like:

1. Interview-me to clarify what the policy needs to address (#1).
2. Step-back to articulate the principle behind your approach (#6).
3. Three-variants of the actual policy proposal (#3).
4. Structured-template for the final draft (#5).
5. Devil’s-advocate against the strongest version (#4).
6. Critique-and-revise for the final polish (#2).

That combines six patterns on one piece of work. Compare the result with a simpler prompt and keep only the stages that improve the output enough to justify their review cost.

A note on what these patterns share

If you look at the ten patterns together, you notice they cluster around three themes:

Patterns that elicit or clarify context — interview-me, contrast-pair, step-back. These help when the original request leaves important goals, examples, or assumptions unstated.

Patterns that structure the model’s response — structured-template, output-template, three-variants, persona. Structure and role constraints can make the requested format and behavior more explicit.

Patterns that ask the model to push back — critique-and-revise, devil’s-advocate, role-play. They request weaknesses or alternative perspectives that a first response may omit.

Many prompt techniques fit into one of these three categories. Use the categories to adapt a pattern when the ten examples do not cover your situation.

Choose by job type

How to internalise them

Reading about patterns does not make them habits. Use them deliberately on real, reviewable work and keep the ones that help. A practical approach:

• Print the list of ten and keep it visible while you work.
• Before sending a non-trivial prompt, pause briefly and ask which pattern, if any, fits.
• If none seem to fit, try the simplest match anyway. The discipline of looking is more important than perfect pattern selection.

With repetition, you may stop needing the printed list and start combining only the patterns that serve the task. Periodically compare the patterned prompt with a simpler baseline so ritual does not replace evidence.

That turns prompting into a more repeatable, testable practice.