Are you ready to maximize Microsoft Copilot’s productivity and unlock its full potential? In this guide, we’ll explore how to effectively leverage Microsoft Copilot across various applications, including Microsoft Office, Dynamics 365, and Bing. Understanding the crucial elements of Persona, Context, and Objective will enable you to craft better prompts that lead to more accurate and relevant responses. By mastering the art of prompt engineering with Copilot, you can supercharge your business operations, streamline workflows, and enhance decision-making processes. Join us as we dive deep into the best practices for utilizing AI-powered assistance and transforming your productivity with Microsoft Copilot!
Copilot is the most powerful productivity tool Microsoft has ever released. And guess what? I can almost guarantee you you’re probably using it wrong.
Let’s talk about how you can take your Copilot answers from zero to hero by learning how to talk to Copilot. At the end, we’re going to talk about the one detrimental mistake you’re making and how to correct it so that you can begin using Copilot.
How to properly talk to Copilot?
Copilot is currently being rolled out across the entire Microsoft suite, including; Office, Dynamics (CRM), Bing, and more. This tool is changing the way people and businesses are getting work done. Unfortunately, few are actually taking full advantage of its capabilities. Copilot, in it’s current state, can feel tricky and have a few pitfalls.
But, the biggest factor that determines Copilot’s output is not your computer. It is not the app you’re using Copilot on. It is not even your database.
It’s you!
What do I mean? Everyone interacts with Copilot through conversation. Therefore, what to write to Copilot is going to have the largest impact on Copilot’s response back to you.
In order to have a good Copilot experience and for it to give you the answer you are looking for, you need to be speaking Copilot’s language. This does not referring to the language you are writing to Copilot and the language it uses to give you an answer. This is referring to how the users of Copilot approach it in a way that is not ideal for Copilot to understand. While the Large Language Models (LLMs) are very advanced, they still cant read your mind.
You interact by providing Copilot with prompts in order to guide Copilot in to providing a valuable response
What is a Prompt?
A prompt is the thing you write, or say, to Copilot.
A prompt, in the context of Copilot, refers to a carefully crafted request that a user submits to the AI-powered code assistant in order to obtain relevant, accurate, and efficient code suggestions. For that reason, the quality of the prompt plays a pivotal role in ensuring that the generated code meets the specific requirements of the task at hand.
To achieve optimal results, it is vital for users to provide a clear, concise, and context-rich description of the problem they are attempting to solve, along with the expected input and output formats. By adhering to these guidelines, users can effectively harness the full potential of Microsoft Copilot, which in turn can significantly enhance productivity and streamline the development process in a professional setting.
What Makes a Good Prompt?
Your prompt is what you write to Copilot and there are three components you need to have in every prompt you write. This is going to allow Copilot to provide you with the best possible answers. The three components are:
- Persona
- Context
- Objective
Persona
The first key component is going to be your persona. This is going to be the person or the role of the author of the prompt. This is essentially telling Copilot who is writing the prompt. It’s important because without a persona, Copilot does not know who it is talking to and it’s not going to produce a quality answer that you’re looking for.
Say you’re a sales manager and you want to write an email to a potential customer that happens to be a big sale. Hence, you want your email to sound different than if it was a sales representative reaching out to that account. Your persona or the sales manager role is going to allow Copilot to write a more desired email and is hopefully going to land that customer.
Furthermore, some examples of personas could be as a customer service agent or as the presenter of an HR presentation. It could even be as the COO of your company to write an email. Thus, the email is going to look and feel different than an HR representative’s email and telling Copilot the persona is going to give it the ability to determine which one you’re looking for.
Context
Now, the second key component to getting a good answer out of Copilot is providing context. Context is going to paint the broader picture for Copilot to determine what it’s actually answering. Context is effectively telling Copilot who its answer is for. Adding context to your prompts is going to be the easiest way to level up the answers Copilot is giving you.
Furthermore, Copilot is only able to use the context of your current message thread. It does not store any information from other conversations happening within your organization or previous conversations you had with Copilot. This means that you’re going to have to give Copilot all the necessary information every time you begin to speak with it. If you ask Copilot for information, it won’t know what you’re talking about without context.
To continue our Sales Manager email scenario, you could include the country that the account lives in, the size of the account, or the industry.
Here’s a secret tip: If Copilot is connected to your Dataverse, you could even call out that account by name and it’s going to have all the information it needs.
Objective
The third and final key component is going to be making sure your prompt includes a clear objective. This is simply the output you are looking for. Hence, the objective is telling Copilot what you are trying to accomplish. It could be to summarize an account, draft and email response, or create a presentation.
How do I level up my Copilot Conversations?
The beauty of Copilot is it is going to look at the entire conversation thread in each response it crafts. If you have a particularly complex ask of Copilot, break your prompt up into several messages. Firstly, write out your persona in detail and finish the message with “Do you Understand?”.
This is going to outline the persona for Copilot going forward. Secondly, define the context of your ask with a new message and conclude it with the same “Do you understand?”. Lastly, provide the objective, your ask, to Copilot.
You can follow this super easy-to-use format below:
Message #1:
I am a Sales Manager for the North Eastern region in the United States for a global manufacturing company. Do you Understand?
Message #2:
I am trying to craft a professional email to send to a large potential customer in my region. The customer is interested in X, Y, and Z. They have reservations over the price and the post-sale care. I want to assure them that we understand their reservations but have confidence in our products ability to add value. This message plays a large role in landing the sale. Do you understand?
Message #3:
Draft me a personal and professional email.
In Conclusion
Making sure you include all three components in every prompt you use with Copilot is going to ensure that your experience is leveled up. If you have noticed that your Copilot experience isn’t great, more often than not, this is because you gave it a bad prompt.
After all, I can be a personal testimony to this. I will get lazy and not speak Copilot’s language by not including one of these three keys. Consequently, I will find that my answer is just not what I was looking for.
Naturally, I’ve been using Copilot to build canvas apps to create and update Power automate flows. I’ve used it to write emails. I’ve used it in my research for different topics. This has given me time to develop an honest review on if Copilot is actually any good. If you’re interested in an honest review of how good Copilot currently is, click on the link here!
Happy Developing!