Additional Rules

The "additional rules" section allows you to customize your copilot's behavior to better meet your specific needs by providing tailored responses, enhanced context, and improved efficiency. By defining clear and specific rules, you can ensure consistent outputs, save time on repetitive tasks, and refine prompts through experimentation. This section also supports particular use cases, making the copilot a more valuable and effective tool for various scenarios. Overall, leveraging the "Additional Rules" section enhances the copilot's functionality and performance, aligning it more closely with your specific requirements.

Here are some tips for creating your own additional rules:

1. "If, then" statements: The additional rules section allows you to create conditional statements that guide the copilot's behavior based on specific conditions. It best understands these instructions in an "if, then" format.
   • Instead of: "Provide a summary of the document."
     - This example does not tell the copilot when it should provide a summary or of what document it should summarize.
   • Use: "If the user's input contains the word 'marketing,' then provide a summary of the marketing strategy."
2. Use brackets for custom inputs: Using brackets around custom inputs lets the copilot know what information to ask the user or obtain from its data source before proceeding. For example, if you want the copilot to write an email, you could include brackets around the recipient's name to prompt the copilot to ask for that information.
   • Instead of: "When writing an email, include the recipient's name."
   • Use: "When writing an email, follow this format: 'Dear [Recipient's Name], I hope this message finds you well.'"
3. Be Clear and Specific: Ensure the instructions are precise and leave little room for interpretation. Using numbers or specific details can help clarify the task.
   • Instead of: "Provide a summary of the marketing strategy."
   • Use: "If summarizing the marketing strategy, only summarize each of the main points in three sentences."
4. Use Contextual Prompts: Provide context to help the copilot understand the task better. For instance, if the copilot is supposed to generate a report, you could include details about the report's purpose, audience, and critical points to cover.
   • Instead of: "When asked, generate a report."
   • Use: "If you are asked to generate a sales report for the marketing team, highlight key performance indicators and trends."
5. Include Examples: Demonstrate the desired behavior by including examples. For example, if the copilot is supposed to write an email, you could provide a sample email that meets your expectations.
   • Instead of: "When asked, write an email to a potential client introducing our new product."
   • Use: "If you are asked to write an email to a potential client, your primary task will be introducing our new product. Follow this example: 'Dear [Client Name], I hope this message finds you well. I am excited to introduce our new product, [Product Name], which offers [key benefits]. I am looking forward to discussing this further. Best regards, [Your Name].'"
     Using brackets around custom inputs lets the copilot know what information it needs to obtain, whether from your data or the user, before proceeding.
6. Leverage Structured Prompts: Use structured formats like bullet points or numbered lists to organize the information. This helps the copilot understand the sequence and importance of each point.
   • Instead of: "If presenting the product, list the features."
   • Use: "If presenting the product, list the features. When listing the features of our product, please expand on the following:
     -User-friendly
     -Real-time collaboration
     -Advanced security features
     -Customizable dashboards"
     
7. Add Supporting Content: Include any additional information that might influence the copilot's output. This could be background information, specific terminology, or any constraints that must be considered.
   • Instead of: "Describe the product and its features when asked."
   • Use: "When you are asked to describe the product, include its main features, target audience, and unique selling points. The product is a cloud-based collaboration tool designed for remote teams."
8. Craft Clear Instructions: Use analogies and descriptive language to help the copilot understand the desired outcome.
   • Instead of: "After greeting the user, give them a product description."
   • Use: "After greeting the user, provide a brief product description for our new cloud-based collaboration tool. Please highlight its ease of use, real-time collaboration features, and advanced security."
9. Provide Cues: Use cues to jumpstart the copilot's output. For example, starting a sentence or providing a partial response can guide the copilot in the right direction.
   • Instead of: "If asked, provide a quarterly sales report."
   • Use: "If an employee needs a sales report, please provide the quarterly report for the specified date range. Begin the report with an introduction that outlines the purpose and scope of the report. For example: 'This report provides an analysis of the quarterly sales performance, focusing on key trends and areas for improvement.'"
10. Experiment and Iterate: We encourage you to experiment with different rules and iterate based on the copilot's responses. This helps in refining the rules to achieve your desired outcomes. Just keep testing and adjusting till your copilot responds how you would like. If you still have problems, use the Send Feedback button at the top of your screen to contact us! We are happy to help!