10 Tips to Write Effective AI Prompts
9 minutes
Prompt engineering is a key skill for getting the best results from AI. Well-crafted prompts lead to clearer, more relevant, and accurate answers from the model.
This guide shares 10 tips to help you create prompts that are simple, organized, and effective.
Quick Reference Guide
| # | Tip Title | Short Description |
|---|---|---|
| 1 | Start with a Minimal Prompt | Begin with a simple instruction and add details gradually. |
| 2 | Be Specific and Clear | Provide detailed and unambiguous instructions to guide the AI. |
| 3 | Provide Context | Define the target audience or purpose to shape the response appropriately. |
| 4 | Structure Your Prompts and Use Clear Separators | Organize your prompt into clearly defined sections for better clarity. |
| 5 | Use Few-Shot Examples | Show sample input-output pairs to demonstrate the desired format or style. |
| 6 | Assign a Role to the AI | Set a persona to help tailor the tone and depth of the AI's response. |
| 7 | Break Complex Tasks into Steps (Chain-of-Thought Reasoning) | Ask for step-by-step reasoning for tackling complex tasks. |
| 8 | Iterate, Validate, and Refine | Continuously improve your prompt based on testing and feedback. |
| 9 | Use Clear Action Verbs | Start with direct verbs to clearly indicate the expected action. |
| 10 | Include Constraints and Specify Format Preferences | Set guidelines like word counts and formats to ensure structured responses. |
1. Start with a Minimal Prompt
Begin your prompt engineering journey with a minimal prompt and gradually add complexity as you learn how the model responds. Starting with a simple instruction allows you to understand the model's baseline behavior without overwhelming it with details. Once you observe the initial results, you can iteratively refine the prompt by adding tone, focus, or specific constraints. This method helps identify which elements are effective and what areas require improvement.
Simple Prompt
Later, after reviewing the output, you might refine it by specifying additional parameters such as style or word limits:
Refined Prompt
By iterating, you can fine-tune prompts to achieve the most relevant and accurate responses.
2. Be Specific and Clear
Precision is essential when crafting prompts. Vague or ambiguous instructions often lead to generic answers that miss the mark. By clearly stating exactly what you expect, you direct the AI to focus on the desired aspects of the topic. Detailed prompts provide boundaries and context that help the model generate more relevant and targeted responses.
❌ Vague Prompt
✅ Specific Prompt
3. Provide Context
Adding context to your prompts significantly improves the model's ability to tailor its response. When you define the target audience or explain the purpose behind your inquiry, the AI can adjust its explanation to meet those needs. Context helps in setting the tone and determining the depth of detail required.
❌ General Prompt
✅ Contextualized Prompt
By clarifying who the explanation is for, you ensure that the response is both appropriate and informative for that specific audience.
4. Structure Your Prompts and Use Clear Separators
A well-organized prompt is more likely to yield a well-organized output. Structuring your prompt into clearly defined sections helps the AI distinguish between different parts of your instruction. You can use sections like "Context," "Task," "Constraints," and "Format" to ensure every aspect of your request is addressed.
❌ Unstructured Prompt
Plan a weekend trip for a family.
✅ Structured Prompt
Context: A family of four is looking for a weekend getaway in the Pacific Northwest.
Task: Create a two-day travel itinerary that balances outdoor activities, local dining experiences, and attractions suitable for children.
Constraints:
- Total budget must not exceed $600.
- Include at least one free activity each day.
- Ensure travel times between activities are realistic.
Format:
Day 1:
Morning: [Activity/Attraction]
Afternoon: [Activity/Attraction]
Evening: [Dining Option]
This version breaks down the request into clear sections that guide the model step by step, ensuring the itinerary is comprehensive, organized, and meets all specified requirements.
5. Use Few-Shot Examples
Providing example input-output pairs, often referred to as few-shot prompting, guides the model by showing it exactly how to transform the input. This method is especially useful when you need a specific format or a particular style of response. By demonstrating the desired outcome, you reduce ambiguity and improve consistency in the results.
One-Shot Prompt
Task: Translate the following sentence from formal to casual tone.
Input: "Could you kindly help me with this issue?"
Output: "Can you help me with this?"
Input: "I would appreciate your assistance with this matter."
Output:
AI Output
"I'd really appreciate your help with this."
Here's another example:
One-Shot Prompt
Task: Extract key factual details from the given news article excerpt. Present the information in the following structured format with these fields: Date, Location, People Involved, and Event Summary.
Input: "On April 15, 2021, a major earthquake struck the city of Los Angeles, California, causing widespread damage and injuring hundreds. Local authorities are assessing the impact, and rescue operations are underway."
Output:
Date: April 15, 2021
Location: Los Angeles, California
People Involved: Local authorities, rescue teams, injured residents
Event Summary: A major earthquake struck Los Angeles, causing widespread damage and numerous injuries, with ongoing rescue and assessment operations.
Input: "On March 30, 2022, a severe winter storm hit New York City, resulting in significant power outages and transportation disruptions. The city's emergency services responded swiftly to assist affected residents and restore order."
Output:
AI Output
Date: March 30, 2022
Location: New York City
People Involved: Emergency services, affected residents
Event Summary: A severe winter storm hit New York City, causing power outages and transportation disruptions. Emergency services responded swiftly to assist residents and restore order.
6. Assign a Role to the AI
Adding a role to the prompt helps the model understand the context and tailor its response. Role prompting is most commonly used to alter the style of AI-generated text. By assigning a role, such as a food critic or a marketing expert, the AI can adjust its tone, language, and the depth of its responses based on the given persona.
Role Prompting
You are a communications specialist. Draft an email to your client advising them about a delay in the delivery schedule due to logistical problems.
AI Output
Subject: Update on Delivery Schedule
Dear [Client's Name],
I hope you’re doing well.
I wanted to inform you about an unexpected delay in the delivery schedule due to logistical challenges we are currently facing. Our team is actively working to resolve the issue, and we are doing everything possible to minimize the impact on your timeline.
At this stage, we anticipate the new estimated delivery date to be [Revised Date]. We will keep you updated on any further developments and notify you immediately if there are any changes.
We sincerely apologize for any inconvenience this may cause and appreciate your patience and understanding. Please let us know if you have any concerns or if there’s anything we can do to support you in the meantime.
Thank you for your continued trust.
Best regards,
[Your Name]
[Your Position]
[Your Company]
[Your Contact Information]
7. Break Complex Tasks into Steps (Chain-of-Thought Reasoning)
Complex tasks benefit from a step-by-step approach. Instructing the model to break down its reasoning into sequential steps ensures detailed and accurate answers. This chain-of-thought method makes the model’s reasoning process transparent and easier to verify.
Chain-of-Thought Prompt
Q: John has 10 apples. He gives away 4 and then receives 5 more. How many apples does he have?
A:
- John starts with 10 apples.
- He gives away 4, so 10 - 4 = 6.
- He then receives 5 more apples, so 6 + 5 = 11. Final Answer: 11
Q: A farmer has 50 apples. In the morning, he sells 1/5 of his apples. In the afternoon, he sells 1/4 of the remaining apples. How many apples does the farmer have left?
8. Iterate, Validate, and Refine
Prompt engineering is an iterative process. Consider your initial output as a draft that you can refine over time. By testing your prompt and then validating the results, you can continuously improve the clarity and focus of your instruction. Refining your prompts based on feedback or errors is key to developing highly effective prompts.
Initial Prompt
Describe the impact of climate change.
AI Output
The impact of climate change is widespread, affecting ecosystems, economies, and human societies in various ways. Here are some of the key consequences:
1. Rising Temperatures
- Global temperatures have increased due to greenhouse gas emissions, leading to more frequent and intense heatwaves.
- Hotter temperatures contribute to droughts, wildfires, and heat-related illnesses.
2. Extreme Weather Events
- More frequent and severe hurricanes, storms, and floods due to changes in atmospheric and oceanic conditions.
- Increased intensity of rainfall in some regions, leading to flash floods and landslides.
3. Sea Level Rise and Coastal Erosion
[Truncated for brevity]
Refined Prompt
Describe the impact of climate change on coastal cities in 200 words. Focus on rising sea levels, economic disruption, and potential mitigation strategies.
AI Output
The Impact of Climate Change on Coastal Cities
Climate change poses a severe threat to coastal cities due to rising sea levels, economic disruption, and the urgent need for mitigation strategies. As global temperatures increase, glaciers and polar ice caps are melting, causing sea levels to rise. This leads to coastal erosion, more frequent flooding, and saltwater intrusion into freshwater supplies, putting millions of residents at risk.
The economic consequences are significant. Flooding damages infrastructure, homes, and businesses, leading to costly repairs and insurance losses. Tourism, a key industry in many coastal cities...
[Truncated for brevity]
This iterative process allows you to gradually narrow down on the most effective prompt for your needs.
9. Use Clear Action Verbs
Starting your instructions with direct action verbs sets the stage for the type of response you expect. Clear action words such as “analyze,” “compare,” or “summarize” provide a specific directive that helps guide the model toward a detailed and focused answer.
❌ General Prompt
Tell me about electric vehicles.
AI Output
Electric vehicles (EVs) are automobiles powered by electricity rather than conventional fossil fuels like gasoline or diesel. They use electric motors and rechargeable batteries, offering a cleaner and more sustainable alternative to internal combustion engine (ICE) vehicles.
Types of Electric Vehicles
- Battery Electric Vehicles (BEVs)
- Fully electric, powered exclusively by batteries.
- No tailpipe emissions.
- Examples: Tesla Model 3, Nissan Leaf, Ford Mustang Mach-E.
[Truncated for brevity]
✅ Specific Prompt
Analyze the advancements in electric vehicle technology over the past decade, focusing on improvements in battery efficiency and charging infrastructure.
AI Output
Over the last decade, electric vehicle (EV) technology has seen remarkable progress, particularly in battery efficiency and charging infrastructure. These advancements have significantly improved EV performance, affordability, and accessibility, making them a viable alternative to internal combustion engine (ICE) vehicles.
- Improvements in Battery Efficiency
- In 2013, most EVs had a range of 100-150 miles per charge, whereas today's EVs commonly exceed 250-400 miles.
- Battery energy density has doubled, reducing the size and weight of batteries while increasing capacity.
- Tesla, Lucid, and Rivian now offer vehicles with 500+ miles of range, a significant leap from early models.
[Truncated for brevity]
These action verbs help shift the focus from a general overview to a more detailed, analytical response.
10. Include Constraints and Specify Format Preferences
Defining constraints like word counts, bullet-point formats, or specific structures can help tailor the AI’s response to your exact requirements. Constraints act as guardrails, ensuring the output remains concise and structured in a way that is easy to interpret and use.
Prompt with Constraints
Summarize the key findings of the latest cybersecurity report in exactly 150 words, using bullet points to list each major insight.
This method ensures that the output not only covers the necessary information but does so in a format that suits your needs.
Conclusion
Creating high-quality prompts is a process of continuous improvement. By following these 10 tips, you can get more precise and useful responses from AI models. Keep experimenting and refining your prompts, and soon you'll see a significant boost in the quality of your AI interactions.
Happy prompting!
Valeriia Kuka
Valeriia Kuka, Head of Content at Learn Prompting, is passionate about making AI and ML accessible. Valeriia previously grew a 60K+ follower AI-focused social media account, earning reposts from Stanford NLP, Amazon Research, Hugging Face, and AI researchers. She has also worked with AI/ML newsletters and global communities with 100K+ members and authored clear and concise explainers and historical articles.