Blog/Artificial Intelligence

How to Write Effective AI Prompts

Copy Link

Toluwani Folayan

March 05, 2024

You're likely familiar with the term ChatGPT. Chances are, you use it every day for your assignments, work, or just about anything else. If you aren't, ChatGPT is one of the many generative AIs you can use for producing text, code, poetry, stories, and much more. However, this article isn't solely about ChatGPT.

We want to discuss AI prompts because the simple truth is, almost everyone knows about these AIs, but not many are fully maximizing their potential. Yes, there are ways you can ask AI for information to get precise results. In fact, most of the time, the outcomes you get from generative AI depend on the types of questions you ask. Ask the right prompts, and you just might get what you need.

It's really not that difficult. It's all about following the tips we will be discussing on how to write effective AI prompts. And if you're interested, you could even get paid as a prompt engineer and make money by just using AI.

Excited? Let’s get started!

Outline

What are AI Prompts?

How to Write Effective AI Prompts

Top AI to Use

What are AI Prompts?

AI prompts are explicit instructions provided to artificial intelligence systems to guide their actions or generate specific outputs. They function as direct commands or inquiries, directing the AI on what tasks to perform or what information to generate.

AI prompts act much like navigating a GPS system. For instance, imagine you're planning a road trip to a remote destination. Without clear directions, you'd likely find yourself lost in unfamiliar roads and dead ends. However, with a GPS device, you simply input your desired destination, and it provides step-by-step guidance, ensuring you reach your destination efficiently.

Similarly, AI prompts serve as the coordinates that steer AI systems toward desired outcomes. They provide clear instructions, enabling users to leverage AI's vast capabilities effectively. Without these prompts, AI systems lack direction and cannot fulfill specific user requests or tasks effectively.

Tips on How to Write Effective AI Prompts

1. Be specific and always add context:

Specificity is key when writing AI prompts. Trust us, every additional information counts and could be that secret ingredient that makes your answer just right!

Stay away from generic prompts like “write a story,” except you want a generic answer. If you want something truly unique, try specifying the type of story you want, the genre, intended target audience, length, tone, language, the type of effect you want readers to feel after reading, and the purpose of the story.

Generic prompt example: Write a story on the importance of self love

Effective AI prompt example: Write a story of 600 words on the importance of self love. The genre should be in prose form and must be in the first person narrative. The tone must be emotional, inspirational, yet humorous. Use simple language. The target audience is adults. I want to post this story on my medium account and I need it to inspire people who are going through difficult times and finding it hard to love themselves. Start with an emotional story that will hook the readers. It could be something like “ I remember the first time I considered committing suicide. My mum had just told me she wished I’d never been born…”

Try these two prompts and compare the results. Which result do you prefer? The first or the second? Remember you can always add more context. The aim is to help the AI understand your prompt.

2. Write like you are talking to a friend

Always have it at the back of your mind that this is a conversation, so approach your prompts in a conversational format. Write like you're talking to a friend. You could even give it a name! This really helps set the tone.

For instance, when conversing with someone, it's natural for them to initially miss your point and require clarification. Similarly, expect the AI to sometimes need guidance or further explanation. Just as you would steer a conversation back on track with a friend, don't hesitate to do the same with your AI prompts. This approach, known as Interactive prompting, encourages a dialogue between you and the AI.

Feel free to ask multi-step questions, where each response informs the next prompt. For instance, ask a question, receive a response, and ask another question based on that response. This interactive approach yields powerful results, much like the dynamic flow of a conversation with a friend. Just as you wouldn't limit yourself to a single question in a conversation, adopt the same conversational style when interacting with the AI.

Don’t be shy to challenge your AI. If you feel the response is wrong, tell it with reasons why you think so. It is a conversation afterall!

3. Do not overload your AI with multiple tasks at once

Focus on single tasks. Once you are done with the first task, then proceed to the next task. Avoid overloaded prompts as this could confuse your AI. They are not built to multitask. Just like tackling one task at a time leads to better focus and productivity in real-life scenarios, the same principle applies to AI interactions.

Presenting the AI with multiple tasks simultaneously overwhelms its processing capabilities and may result in inaccurate responses. However, by breaking down complex requests into individual tasks, you provide clarity and direction to the AI, enabling it to generate more accurate and relevant responses. So, take it one step at a time, and allow your AI to concentrate on each task sequentially for optimal performance

4. Give it an idea of what you want your result to look like

It's all about painting a picture for your AI. You have the vision, so you must paint it so the AI sees it too. Generative AI can produce a wide array of outputs, ranging from code and stories to reports, summaries, dialogue, and beyond. To ensure the AI understands your desired result, be specific about the type of output you want.

After describing what you're looking for, try incorporating phrases like "Present this in the form of…" followed by your preferred output format. Alternatively, start your prompt with "Create a [format of output] about / that contains…" This approach provides the AI with a clear image of the final product you're envisioning, leading to more accurate and satisfactory results.

Generic AI prompt example: Generate customer service responses for a food delivery company.

Effective AI prompt example: Create a set of customer service responses tailored for addressing inquiries related to food delivery orders for our company. Present this in the form of clear and helpful responses, offering assistance with order tracking, delivery time estimates, resolving issues with missing or incorrect items, and providing information on promotional offers and discounts.

5. Add your do’s and don’ts

Specify your preferences and requirements clearly in your AI prompts. This is because providing clear instructions about what you do and don’t want in your response helps the AI understand your expectations and generate more relevant outputs.

Additionally, avoid assuming that the AI will automatically understand your preferences without explicit instructions. Without clear guidance, the AI may include elements that are not aligned with your needs, resulting in less accurate or satisfactory responses.

Generic AI prompt example:Generate an unique proposal for a new marketing campaign for an energy drink.

Effective AI prompt example: Create a proposal for a new marketing campaign targeting young adults aged 18-25, promoting a new energy drink product. Do include innovative social media strategies, influencer partnerships, and engaging visual content. Don't include traditional print advertising methods or messaging targeting older demographics. Ensure the proposal emphasizes the product's energy-boosting benefits and aligns with the brand's adventurous and youthful image. Also avoid sounding too salesy or promotional. It must be natural, unique and make people interested in the energy drink.

6. Familiarize yourself with your AI’s compatibilities

While it might be tempting to think, "It's AI! What can't it do?" the reality is that each AI has its limitations. For example, ChatGPT cannot generate images or provide direct links to articles. Understanding these limitations is crucial to avoid requesting tasks beyond the AI's capabilities, which could result in inaccurate or misleading information.

Furthermore, different AI systems possess varying strengths and weaknesses, necessitating a tailored approach. Therefore, take the time to research and understand the tool's functionalities before crafting your prompts. Identify its strengths and weaknesses to optimize your use of the AI. This proactive approach ensures that you leverage the AI's capabilities effectively and avoid potential pitfalls.

7. Rephrase questions

If your AI isn't providing the desired responses, consider rephrasing your questions in various tones and formats. Experiment with different wording and approaches until you obtain the desired outcome. This process involves persistence and flexibility in adapting your communication style to elicit the desired information from the AI.

For instance, let's say you're using AI to help with a school assignment on climate change. Your initial prompt, "What are the causes of climate change?" yields vague or unrelated results. Instead of giving up, you rephrase your question in different ways and tones:

"Could you provide a list of factors contributing to climate change?"

"What are the main reasons behind global warming?"

"Can you explain the human activities that impact the Earth's climate?"

"I'm researching climate change. How do human actions affect the environment?"

"Could you outline the primary drivers of climate change, please?"

By trying various approaches, you increase the chances of obtaining the specific information you need for your assignment.

8. Open a new page if you are not getting your desired results

Consider starting a new chat session if you're not getting the desired results. Sometimes, the AI may get stuck in a pattern based on the format of your previous questions. However, opening a new chat or tab allows the AI to provide a fresh response or approach to your query.

Additionally, if you've been asking questions in the same chat and still not getting the information you need, initiating a new chat can be beneficial. This approach can help alleviate any confusion or difficulty the AI may encounter in keeping track of the conversation thread.

Essentially, by starting anew, you give the AI an opportunity to reset and provide more relevant and accurate responses.

9. Try asking multiple AIs the same question

Experiment with asking the same question to multiple AI models. This is because different types of AI, such as ChatGPT, Bard, Claude, Synthesia, GitHub Copilot, and many others may interpret and generate responses differently despite being tasked with the same query. However by opening separate tabs and posing identical prompts to each AI, you can observe how their answers vary.

For instance, if you ask, "What are the effects of climate change on biodiversity?" to ChatGPT and Bard, you may receive distinct responses based on their respective training data and algorithms. ChatGPT might provide a broad overview, while Bard may offer a more detailed scientific explanation.

Basically, analyzing the responses from multiple AI models allows you to compare their insights and perspectives. You can then decide which response best aligns with your needs or combine elements from each to create a comprehensive understanding of the topic. This approach leverages the strengths of different AI models and enhances the quality of information obtained.

10. Ask for examples

Asking for examples is another effective strategy when interacting with AI. Anytime you request examples, you prompt the AI to provide concrete instances or illustrations related to your query, which can enhance your understanding of the topic.

For instance, you can try saying, "Can you provide examples of how this concept applies in real-world situations?" This prompts the AI to generate specific instances or scenarios that demonstrate the application of the concept you're exploring.

Also, by asking for examples, you not only gain clarity on abstract concepts but deepen your comprehension by contextualizing them in real-world contexts. This approach can enrich your learning experience and aid in problem-solving or decision-making processes.

11. Ask it to repeat your original requests

Another effective approach to ensure clarity when interacting with AI is to ask it to recap your original request. This helps confirm whether the AI has understood your query correctly and is on the right track to providing a relevant response. You can prompt the AI by saying, "Can you try summarizing my original request, I just want to be sure you understand my query?" This request encourages the AI to restate your inquiry in its own words, allowing you to verify its understanding and refine your request if necessary. It serves as a simple yet effective method to enhance communication and ensure that the AI addresses your needs effectively.

Top Generative AI to Use

1. ChatGPT:

Developed by OpenAI, ChatGPT is a state-of-the-art language model designed for conversational interactions. It uses deep learning techniques to generate human-like responses to text inputs. People can use ChatGPT for a variety of purposes, including answering questions, providing recommendations, generating creative content such as stories or poems, and even engaging in casual conversation. Its versatility, natural language understanding, and ability to produce coherent and contextually relevant responses make it a popular choice for various applications.

2. Bard:

Bard is another generative AI developed by OpenAI, specializing in creative writing tasks such as generating stories, poems, and narratives. It is trained on a diverse dataset of literary works to mimic the style and tone of human authors. Users can also leverage Bard to create original content for storytelling, entertainment, or artistic expression. Its ability to generate high-quality and engaging literary pieces makes it a valuable tool for writers, educators, and content creators.

3. Perplexity:

Perplexity is a metric commonly used to evaluate the performance of language models, including ChatGPT and Bard. It measures how well a language model predicts the next word in a sequence of text. Lower perplexity scores indicate that the model can more accurately predict the next word, implying better performance in understanding and generating text. Perplexity is crucial for assessing the quality and effectiveness of generative AI models in various natural language processing tasks.

4. GitHub Copilot:

GitHub Copilot is an AI-powered code completion tool developed by GitHub in collaboration with OpenAI. It assists developers by suggesting code snippets and completing repetitive tasks based on context and natural language descriptions. Developers can also use GitHub Copilot to write code more efficiently, reduce errors, and explore new programming techniques. Its integration with popular code editors and ability to understand code patterns make it a valuable resource for software development.

5. GPT-4:

GPT-4 is a the successor to GPT-3, a widely used language model developed by OpenAI. Similar to its predecessors, GPT-4 can be used for various text-based applications, including conversational agents, content generation, language translation, and more. Its potential improvements in accuracy, coherence, and context awareness make it an even more powerful tool for natural language processing tasks. It is a paid premium and more improved version of gpt 3 which is free.

6. AlphaCode:

AlphaCode is a generative AI model developed by OpenAI focused on code generation and programming tasks. It is designed to understand natural language descriptions of programming problems and provide code solutions automatically. AlphaCode aims to assist developers in writing code more efficiently, debugging programs, and exploring new programming paradigms. Its ability to understand context and generate syntactically correct code makes it a valuable resource for software development projects.

Conclusion

Maximize the benefits of Artificial Intelligence by putting these proven tips into practice. With these simple yet effective strategies, you'll see a noticeable improvement in how AI responds to your queries. Remember, by being specific, clear, and contextual in your prompts, you'll guide AI towards generating more accurate and relevant answers. Don't settle for mediocre results— start refining your prompts now!