Skip to main content

​
Craft better prompts

Well-designed prompts help you get more precise and useful responses from your AI agents.

​
General prompt patterns

Basic structure

An effective prompt typically includes:
[Task/Request] + [Specifics] + [Format/Output preferences]
Example: “Summarize this quarterly report, focusing on revenue trends and customer acquisition metrics. Present your analysis with bullet points for key findings and a short paragraph for implications.”
Provide relevant background information:
[Situation/Context] + [Task/Request] + [Purpose]
Example: “I’m preparing for a job interview at a financial technology company for a product manager role. Help me prepare 5-7 questions I could ask the interviewer that demonstrate my knowledge of the fintech industry and my interest in their business model.”
Specify how the agent should approach the task:
Act as [role/expert] and [task/request]
Example: “Act as an experienced copywriter specializing in email campaigns and review this draft newsletter. Suggest improvements to increase open rates and engagement.”
Request detailed walkthrough:
Guide me through [process] step by step, explaining [specific aspects]
Example: “Guide me through creating a basic data visualization in Python step by step, explaining the purpose of each line of code. I’m a beginner with programming but understand basic concepts.”

​
Task-specific examples

Topic exploration:
Provide an overview of [topic], covering:
1. Key concepts and definitions
2. Historical development
3. Current state of research
4. Major debates or unsolved questions
5. Practical applications
Comparative analysis:
Compare and contrast [A] and [B] in terms of:
- Core principles
- Strengths and limitations
- Use cases
- Implementation considerations
Academic literature:
Summarize the current research on [topic], highlighting:
- Major findings from the last 3-5 years
- Methodological approaches
- Gaps in the literature
- Directions for future research

​
Format-specific prompts

Data analysis

Analyze this dataset of [description]:
[data or file]

Please provide:
1. Summary statistics for key variables
2. Identification of notable patterns or outliers
3. Visualization recommendations
4. Potential correlations worth exploring
5. Limitations of this analysis

Code generation

Write a function in [language] that:
- Takes inputs: [parameters]
- Performs: [functionality]
- Returns: [output]
- Handles these edge cases: [exceptions]

Include comments explaining the approach and any non-obvious parts.

Learning materials

Create learning materials about [topic] for [audience level]:

1. Key concept explanation with analogies
2. Step-by-step tutorial for [specific task]
3. Common misconceptions and clarifications
4. Practice exercises with increasing difficulty
5. Application examples in real-world contexts

Feedback formulation

Help me provide constructive feedback on this [work]:
[content]

Structure the feedback as:
- Strengths: What works well
- Areas for improvement: Specific suggestions
- Questions: Points to clarify
- Next steps: Actionable recommendations

The tone should be supportive but honest.

​
Refining responses

If you’re not satisfied with an initial response, try these follow-up prompts:
When you needTry asking
More depth”Could you expand on [specific point] with more detail?”
Simplification”Please explain that more simply, as if to someone new to the subject.”
Different perspective”How would someone with [different viewpoint/background] approach this?”
Practical application”Could you provide specific examples of how to apply this in [context]?”
Evidence”What data or research supports these conclusions?”
Alternatives”What are some alternative approaches or interpretations?”
Instead of starting over with a new prompt, build on the conversation by refining your request iteratively.

​
Next steps

Productivity Hacks

Discover techniques to streamline your agent workflows