The Science of Prompts: How to Write Effective Prompts

In recent years, the pace and volume of work have increased: on a typical workday, for example, Teams users attend three times as many meetings per week as they did in 2020, and on Windows, some people use 11 apps in a single day to get their work done.

This is precisely why the assistance of an artificial intelligence tool like Copilot tangibly improve people’s work lives, effectively enhancing human capabilities and ushering in a“new era of productivity.”

But what exactly is it? Copilot a personal assistant integrated into the M365 ecosystem, Microsoft Edge, and Windows. This innovation promises to deliver a wide range of benefits: from boosting daily efficiency in programs like Excel or PowerPoint to enhancing the user experience. In short, it’s like having a new colleague who’s always ready to help us!

However, an innovation like Copilot just represent a technological paradigm shift—it also involves cultural transformation and a new mindset; in short, a true harmonization between people and technology. This is precisely why the tool must be adopted correctly:“AI can truly elevate your organization’s potential and redefine the limits of what can be done. But collaborating with AI is more than just pressing buttons. It’s about asking the right questions,” says Jaime Teevan, Chief Scientist & Technical Fellow at Microsoft.

One of the hottest topics in the field of AI adoption is undoubtedly that of“asking the right questions”of artificial intelligence—in other words, writing prompts correctly.
A prompt is a piece of text written in natural language that is used to ask artificial intelligence to generate a specific result.

‍

How to Write Effective Prompts

First of all, there’s a general rule: the more detail you provide in your prompts, the better the response will be. If you ask Copilot or another AI) a very vague or very simple question, it will give you a very general answer. To get more detail and precision, your question needs to be more specific.

Let’s take an example: if you ask Copilot, “What did I miss yesterday?”, it will offer you a series of summaries of emails and chats in which you were included. For greater precision, you need to be more specific and detailed in your initial interaction with Copilot—for example, by writing: “What tasks was I supposed to have completed by yesterday for client XYZ?” This will provide more accurate results that summarize information from meetings, email discussions, chats, and comments in documents.

"In traditional search, people have gotten used to being very concise," says Tara Roth, Corporate Vice President of Microsoft 365 Customer Success Engineering. "We use just a few keywords and get a lot of links, but you have to do a lot of the work yourself to make sense of them. With prompts, you can be more descriptive about what you want, and you can get more relevant and accurate answers."

Another tip for a successful Copilot prompt Copilot to break the process down into 4 steps:

1. Start with the end in mind: in other words, explain what you want Copilot to Copilot ;
2. Context: Provide the appropriate context, details, and additional information;
3. Parameters: Specify any parameters Copilot use (specific dates, documents, emails).
4. Presentation: Describe how you want Copilot to Copilot its response

In any case, you still need to be creative and experiment with different prompts to get answers that meet your needs. For example, you can try specifying the tone of voice ( neutral, informal, professional) or providing guidance on the type of language to use—such as language that a non-technical person could understand. Analogies and even historical allegories can also be useful ways to help you process the information (e.g. , what is a moment in history I can use to explain the central message in this document?).

Some concrete examples of good prompting
‍

HR Case Study

‍
Once a week, to stay informed and prepare for meetings with global HR partners, I ask Copilot:

· Are human resources trends in the United States in September 2023 different from those in Germany?

· Show me the human resources data for all of our global offices for the last three months of 2023.

· Who are the most authoritative voices discussing these challenges?

Try rephrasing the sentence.

Here are some examples of how paraphrasing a prompt can influence the output generated by Copilot. Changing the style can help you achieve different results depending on your needs.

· What does the word "epistemology" mean?

Epistemology is a branch of philosophy that deals with the origin, nature, and validity of knowledge. It focuses on how we acquire knowledge, how we can be sure that it is accurate, and what value knowledge has. In other words, it is the study of the process of knowledge.

· Can you explain to me what "epistemology" means?

Yes, of course! Epistemology deals with knowledge and focuses on questions about how we know what we know. In simple terms, it seeks to answer questions such as “How do we know what we know?” and “Can we really know anything for certain?”

· Can you give me a clear definition of "epistemology"?

Epistemology is a branch of philosophy that deals with the nature, origin, and validity of knowledge. Essentially, it seeks to answer questions such as “How can we know things?” and “To what extent can we be certain of what we know?” This field plays a fundamental role in analyzing the processes of acquiring and validating knowledge.

Using Tables and Diagrams

In some cases, you may need structured data presented in tables, charts, or diagrams. Copilot generate these visualizations based on your requests.

Here is an example of a prompt to generate a data table: "Create a table showing a comparison of sales for products A, B, and C for the first six months of 2023."

With a prompt like this, Copilot a table with the requested data.

Another example of a prompt could be to generate a chart: "Create a bar chart showing our company's revenue growth over the past three years."

Copilot create a bar chart showing revenue growth over the specified period.

The use of tables, diagrams, and charts can make it easier to understand and visualize data, making the presentation of information more efficient.

Be creative!

Creativity in prompts can lead to more surprising and thought-provoking responses. Ask Copilot generate ideas, tell stories, or explore future scenarios. Here are some examples of creative prompts:

· "Imagine a world where cars fly. How would that change people's daily lives?"

· "Tell me a short story about the colonization of Mars by humans."

· "Explore the potential ethical implications of artificial intelligence in healthcare."

These prompts stimulate Copilot creativity Copilot can lead to interesting and in-depth responses on complex and visionary topics.

Always use context to optimize your responses

Providing context in your prompts can Copilot more targeted and relevant responses. You can specify the time, location, industry, or other relevant details. Here are some examples:

· "What have been the most recent trends in the renewable energy sector in Europe over the past two years?"

· "Show me the statistics on the adoption of e-learning in elementary schools in the United States during the pandemic."

· "Please provide me with a summary of our company's progress in the second quarter of 2023, with a particular focus on our results in China."

Presentation of Responses and Human-AI Collaboration

The answers that an AI like Copilot cannot be taken at face value but must be reworked, and this is precisely where our “human” ability comes into play 100%. Copilot a support tool, but human judgment and discernment remain essential. Make sure the AI complements your own intelligence, not replaces it, and before using Copilot responses, Copilot ensure they are accurate and reliable. If necessary, conduct additional research to validate the information.

Here are additional guidelines for submitting the answers provided by the AI:

Summary and Organization: If you receive detailed responses, you can summarize and organize them into a more digestible format. For example, you can extract key points or create bulleted lists to highlight the most relevant information.

Data Visualization: If Copilot generated tables, charts, or diagrams, you can embed them in your presentations or documents. Make sure they are legible and well-formatted.

Customization: You can tailor Copilot responses Copilot your needs. Edit the language, structure, or details so they align with your audience’s expectations.

‍

‍

In conclusion, AI is revolutionizing the way we work, but to reap its full benefits, it is essential to learn how to work with it. The future of work will be a blend of human creativity and artificial intelligence. By learning to fully harness the potential of AI, you can enhance your decision-making skills, solve complex problems, and achieve greater efficiency.

Copilot vast potential—let's learn how to use it effectively so we can decide how best to implement it in our company!