Proven Techniques to Write Successful Copilot Prompts

As your organization begins to integrate Microsoft Copilot into daily workflows, it's important to master the art of prompt engineering.

What is a Prompt?

A prompt is essentially a question or instruction you give to Copilot to guide its responses. Think of it as a key to a treasure chest of capabilities. The more precise and detailed you structure prompts, the more accurate and useful the response will be.

Simply type your query or command, and Copilot will respond with relevant information, suggestions, or solutions. When Copilot receives a prompt, it uses a large language model based on the GPT-4 architecture to create a response.

Using Copilot prompts helps employees and business leaders work better and make smarter decisions. This creates a more productive work environment. In this blog, we provide you with expert prompting techniques and successful prompt structure examples for getting the most out of Copilot!

4 Components of a Successful Prompt Structure

To write an effective prompt, you need to ensure that it consists of these 4 key elements:

• Goal: Clearly defining the goal helps Copilot understand the exact nature of the information you're seeking. For example, if you're investigating a potential security breach, your goal might be to identify the source of the breach and the extent of the damage.
• Context: Providing context ensures that Copilot tailors its response to your specific situation. For instance, if you're preparing a report for senior management, you might need a high-level summary of the incident and its implications.
• Expectations: Setting clear expectations helps Copilot deliver information in the most useful format. For example, you might need a detailed technical analysis for your IT team or a concise executive summary for non-technical stakeholders.
• Source: Specifying sources ensures that Copilot uses the most relevant and accurate data. This could include internal logs, threat intelligence feeds, or specific plugins that provide additional context or analysis.

The visual below shows an example of how the 4 components are featured in a prompt structure.

Give Copilot a Detailed Persona

Provide clear instructions about who you want Copilot to represent in the prompt. This will change how Copilot responds, including its style and focus.

A well-defined persona ensures consistency in tone, style, and approach, which builds trust and reliability with users. Additionally, a clear persona helps Copilot adapt to different communication styles and preferences, offering a more personalized experience. 

It also ensures that responses are accurate, respectful, and safe, minimizing the risk of false information.Overall, a detailed persona maximizes Copilot's potential by making interactions more effective, enjoyable, and trustworthy.

Order of Your Prompts Matter

The sequence in which you present information in your prompt significantly influences the quality and relevance of the response. It can shift the emphasis Copilot places on different parts of your prompt.

When generating responses, Copilot tends to prioritize instructions or details presented later in the prompt. This behavior can be particularly helpful when you're dealing with complex tasks that require specific context or examples. By carefully arranging your prompt, you can guide Copilot to produce outputs that better align with your goals. 

When using Copilot in Microsoft Word, include any specific files or references at the end of your prompt. This will help improve how it works with your content.

To illustrate this point, let’s compare two prompts aimed at achieving the same goal but structured differently:

Using these prompts, you’ll likely notice differences in Copilot’s output:

• Prompt 1 places more emphasis on following the instruction ("craft a blog post") while potentially treating the context and example as secondary details.
• Prompt 2 integrates the context and example more deeply into the structure of the response since they are presented first, leading Copilot to treat them as foundational information.

Iterate and Regenerate

When you use Copilot or any generative AI tool, keep in mind that the first result is usually not the final or best version. Often, the magic happens through iteration. By refining and tweaking your prompts, you can significantly improve the relevance, depth, and quality of the responses.

If the response doesn’t meet your needs initially, don’t stop there—revise the prompt and try again. Below is a step by step example of how to evolve a prompt to achieve a more effective response:

• Round 1: Write about advancements in renewable energy technologies.
• Round 2: Write a technical article discussing recent advancements in renewable energy technologies, focusing on solar and wind energy, for an audience of engineers and environmental scientists.
• Round 3: Compose a detailed technical article for engineers and environmental scientists, exploring recent advancements and innovations in renewable energy technologies, with a focus on solar and wind energy. Discuss the scientific principles behind these technologies, their efficiency improvements, and real-world applications. Refer to recent research papers, patents, and industry reports for accurate and up-to-date information.

Conclusion

Ultimately, mastering the art of writing effective prompts is crucial for getting specific and accurate responses from Copilot.  

To help Copilot give you accurate and useful responses, focus on four key components: goal, context, expectations, and source. Understanding and using these components will guide your interactions effectively.  

Remember, clearly defining your audience and the order of your prompts is key to getting better output. Iteration and refinement are also crucial steps in achieving the best results.