Prompting is becoming an essential skill for analysts working with generative AI tools like Microsoft Copilot. But prompt writing in Excel isn’t quite like chatting with ChatGPT. Excel’s strengths are all about fast analyses, quick prototyping, and immediate visual insights. Prompts here need to be concise, clear, and focused on data already present in your workbook.
In this post, we’ll break down what makes prompting in Excel a bit different from more general AI tools and walk through a few frameworks that help you get better results.
Why Excel prompting is different
Unlike general AI chat tools, Excel prompts should be short, direct, and data-centric. Excel is best used for quickly testing ideas, exploring data, or generating insights on the spot. That means effective prompts should:
- Reference specific Excel Tables or clearly structured datasets.
- Focus on immediate and straightforward tasks.
- Avoid overly complicated instructions or external file references.
Keeping prompts targeted helps Copilot quickly deliver precise and actionable results.
Microsoft’s recommended prompt format
Microsoft suggests this simple framework to help Excel users get the best results:
- Goal: Clearly state what you want.
- Context: Explain briefly why you need it.
- Expectations: Specify your preferred output (table, chart, formula).
- Source: Point directly to your Excel worksheet or table.
To learn more about this framework, you can visit an earlier post:
While this structured approach is useful for clearly defined, straightforward tasks, Excel users often need flexibility and speed. For rapid prototyping or iterative exploration, the detailed context or specific output expectations might sometimes be too restrictive. Analysts might also find that explicitly defining context and expectations slows down their workflow for simpler tasks.
This is why it’s helpful to know additional, more streamlined frameworks tailored to different scenarios and levels of detail.
Helpful prompt frameworks for Excel
Here are three easy-to-use frameworks to help guide your interactions with Excel Copilot:
1. RISEN Framework
RISEN is ideal for routine or recurring tasks, providing structured clarity to standard analyses and reporting.
- Role: Define your analytical role.
- Instructions: Clearly explain the needed analysis.
- Steps: Provide concise steps for Copilot.
- End Goal: State clearly what Copilot should deliver.
- Narrow: Specify your Excel data.
Example (Sales):
“As a sales analyst (Role), summarize quarterly revenue by product category (Instructions). Calculate totals, highlight top-performing categories, and identify any declining products (Steps). Provide a concise summary table with clear conditional formatting (End Goal), using data from ‘ProductSales_Q1’ on ‘SalesReport’ (Narrow).”
Example (HR):
“Acting as an HR analyst, review employee survey results. Identify common issues, summarize key findings clearly, and highlight critical feedback points. Present a summary list, focusing only on ‘EmployeeFeedback2024’ data.”
RISEN ensures each analytical step is clearly communicated, helping Copilot deliver precise, structured insights efficiently.
2. STAR Framework
The STAR framework excels at clearly defined, targeted business scenarios where the problem and required analysis are straightforward and specific.
- Situation: Clearly describe the current issue.
- Task: Specify the analytical task.
- Action: Directly instruct Copilot how to proceed.
- Result: Clearly define your expected outcome.
Example (Finance):
“Budget overruns occurred in multiple departments (Situation). Identify which departments exceeded their Q1 budgets (Task). Calculate percentage over budget and pinpoint major expense categories (Action). Provide a brief report listing the departments, overspend percentages, and suggested cost-saving actions (Result), using data from the ‘Budget_vs_Actual_Q1’ worksheet.”
Example (Marketing):
“Our recent email campaign had mixed results. Determine open rates and conversions by email type (Task). Calculate and compare performance metrics (Action). Provide a clear summary highlighting best-performing emails and suggested improvements (Result), referencing the ‘EmailCampaignResults’ table.”
STAR makes it easy to clarify your analytical objective, guiding Copilot to deliver exactly what you need quickly and clearly.
3. SOAR Framework
The SOAR framework is designed for detailed, complex analyses that require a systematic, step-by-step approach to achieve a clear and actionable insight.
- Situation: Set the analytical context.
- Objective: Define your analytical goal.
- Action: Specify detailed analysis steps.
- Result: Clearly describe the desired insights.
Example (Operations):
“Production delays increased this quarter (Situation). Our objective is to identify key delay factors and recommend solutions (Objective). Analyze delay data by product line and month, rank delay causes, and identify actionable improvements (Action). Provide a concise report outlining the main delay factors and recommendations to address them (Result), using data from ‘ProductionDelays_Q1’.”
Example (Sales):
“Regional sales targets were missed last month (Situation). Identify which regions fell short and why (Objective). Calculate sales vs. targets, rank regions by performance, and analyze contributing factors (Action). Summarize findings and recommended strategies to improve regional performance (Result), focusing on data from ‘RegionalSalesTargets’.”
SOAR provides clear, structured reasoning steps, helping Copilot systematically analyze data and present actionable conclusions.
Quick tips for Excel prompts
Here are practical, confidence-building tips to ensure effective Excel prompts:
- Always use Excel Tables: Structured tables help Copilot quickly interpret your request accurately.
- Explicitly request Excel features: Clearly specify tools like PivotTables, charts, or conditional formatting for precise outcomes.
- Keep instructions straightforward: Simple prompts minimize confusion, ensuring fast, clear responses.
- Start small and simple: Gain confidence by starting with simple analyses and gradually move to more complex tasks.
- Review Copilot’s output: Always briefly verify results for accuracy, ensuring confidence in your analysis.
- Use iterative prompting: It’s okay to refine your prompt based on initial results—Copilot thrives with iterative feedback.
Excel and Copilot: Powerful Partners
Excel and generative AI tools like Copilot naturally complement each other, excelling together in environments requiring quick, clear, actionable insights. By mastering targeted prompts, you can significantly enhance your Excel workflows and decision-making processes.
Curious about applying these frameworks in your organization? If you have questions or want support integrating Excel Copilot effectively with your team’s workflow, please get in touch. I’d love to help you unlock the full potential of Excel and Copilot.
Leave a Reply