Prompt Engineering Patterns
You don't need to be a developer to write great AI prompts. The best prompts follow proven patterns — simple structures that dramatically improve the quality of what AI produces. Here are the patterns that matter most, with examples you can use today.
Pattern 1: Role Prompting
The simplest and most impactful pattern. You tell the AI who it isbefore telling it what to do. This changes the AI's vocabulary, tone, depth, and perspective.
Example: Getting Financial Advice
Without a role:
Result: Generic overview that could come from any blog post.
With a role:
Result: Specific frameworks, tier suggestions, common mistakes to avoid.
When to use Role Prompting
- •Any time you want domain-specific expertise
- •When tone and perspective matter (e.g., writing for a specific audience)
- •As the foundation for every other pattern — always start with a role
Common mistake:Using vague roles like “expert” or “professional.” The more specific the role, the better the output. “Senior content strategist who specializes in B2B email campaigns” beats “marketing expert” every time. For real examples in two common functions, compare our AI marketing assistant guide with these developer prompt patterns.
Pattern 2: Chain of Thought
Instead of asking the AI for a final answer directly, you ask it to think through the problem step by step. This dramatically improves accuracy for anything involving analysis, reasoning, or complex decisions.
Example: Evaluating a Business Decision
Direct approach (often shallow):
Chain of Thought approach (much deeper):
When to use Chain of Thought
- •Complex decisions with multiple factors to weigh
- •Math, logic, or analytical problems
- •When you want to see the reasoning, not just the conclusion
Common mistake:Skipping this pattern for “simple” questions. Many questions that seem simple actually benefit from step-by-step reasoning. When in doubt, add “Think through this step by step” to your prompt.
Pattern 3: Few-Shot (Show, Don't Tell)
Instead of describing what you want, you show the AI examples of good output. This is the single most effective way to control format, tone, length, and quality.
Example: Writing Product Descriptions
“Write product descriptions for our e-commerce store. Here are two examples of the style and format I want:
Merino Wool Crewneck
Warm without the bulk. Our merino crewneck is soft enough for all-day wear and tough enough for weekend hikes. Machine washable. Runs true to size.
Canvas Weekend Bag
Your go-to for two-night trips. Water-resistant canvas, brass hardware, and a shoe compartment that actually works. Fits overhead on any regional flight.
Now write one for our new titanium water bottle (32oz, double-walled, keeps drinks cold 24hrs).”
When to use Few-Shot
- •When tone, format, or style is hard to describe but easy to show
- •Generating content that matches existing brand standards
- •Any repeatable task where consistency matters
Common mistake: Providing only one example. Two or three examples help the AI distinguish your actual pattern from coincidence. If your examples share a structure, the AI will follow it reliably.
Pattern 4: Constraint Setting
AI tends toward long, generic answers when left unconstrained. Adding explicit constraints — length, format, tone, things to avoid — focuses the output and saves you editing time.
Useful Constraints to Set
Length
"Keep it under 200 words"
Format
"Use bullet points, not paragraphs"
Tone
"Write conversationally, as if to a friend"
Audience
"For a non-technical CEO"
Exclusions
"Don't use jargon or buzzwords"
Structure
"Start with the recommendation, then the reasoning"
Pro tip: The best prompts combine all four patterns. Start with a role, add chain-of-thought reasoning, provide examples via few-shot, and set constraints on the output. This is exactly what professional skill files do — they bake these patterns into a reusable configuration so you get great results every time without rebuilding the prompt from scratch.
Putting It All Together
Here's what a prompt looks like when you combine all four patterns:
[Role] “You are a senior content marketer who specializes in LinkedIn thought leadership for B2B SaaS founders.
[Chain of Thought] Think through what makes a LinkedIn post go viral in the B2B space: hook, value, formatting, and CTA.
[Few-Shot] Here's an example of a post that performed well for us: [paste example]
[Constraints] Write a post about our new AI feature. Keep it under 150 words. Use line breaks for readability. End with a question to drive engagement. No hashtags.”
You don't need to use all four patterns every time. But knowing them gives you a toolkit to reach for when the AI's default output isn't good enough. And if you'd rather skip the prompt engineering entirely, our Marketing Agent SkillPack packages these patterns into ready-to-use configurations — role, examples, constraints, and workflows all included.
Key Takeaways
- Role Prompting gives AI a specific perspective and expertise — always start here.
- Chain of Thought asks the AI to reason step by step, dramatically improving analysis and decisions.
- Few-Shot examples show the AI what good output looks like — two or three examples beat a page of instructions.
- Constraints (length, format, tone, exclusions) prevent AI from defaulting to long, generic responses.
Unlock the complete Learning Hub
Get full access to all training content for just $5
Learn how to build expert-level AI skills, master prompt engineering, and 10x your productivity
Unlock for $5Already purchased? Restore access
One-time payment • Lifetime access