Advanced Analysis: Multi-Step Analytical Workflows

1. Advanced Analysis: Multi-Step Analytical Workflows

Until now, we've mostly operated within Excel's typical workflows, we've simply accelerated the process and made it more joyful.

2. When standard tools fall short

But what if we want to push these capabilities to understand how spend, revenue, units, and region all correlate? This sort of analysis takes multiple steps, with calculations, visualizations, and careful validation of each step. There's often a lot of manual work here, and many people switch to coding tools to avoid this. Or imagine we have 500 free-text survey responses where we need to identify the top themes—no Excel formula can do that. These scenarios need multi-step exploration, and Copilot can help us push Excel to its limits.

3. How Copilot performs multi-step analysis

If we gave Copilot a prompt like: "Investigate the correlation between the variables in the workbook." Copilot reasons through the data available, plans the series of steps it will take to conduct the analysis. Then, if Agent mode is enabled, it will start editing your workbook to create any supporting tables and visualizations to help you make a conclusion. Because everything is done step-by-step, you can inspect every step of the analysis, leaving nothing to chance. And remember, if you don't want Copilot to edit your workbook directly, you can just switch agent mode off. Let's test this out!

4. Example: Investigating Correlation

We open Copilot with Agent Mode enabled and ask: "Investigate the correlation between salary, job title, and region in the dataset provided." Copilot begins by planning its analysis, and may open an interface to help you track this. Quite often, Copilot will create a separate analysis tab so it doesn't clutter-up your existing workbook. Once it's finished, we get a summary of it findings. To understand how it got here more clearly, we can view the tables and visuals it used to come to its conclusions. Notice that it didn't just create formulas and tables all over the place - there's a clear structure to the work it did, and this could easily be shared with an analyst for verification. Finally, it will often output an interpretation table to help you make a conclusion. Now let's tackle something formulas can't do: summarizing free-text.

5. Example: Summarizing free-text into themes

In a different workbook, we have a table of user feedback. We can ask Copilot: "Summarize the top improvement themes from the feedback column." Copilot reads hundreds of comments and groups them into themes—maybe "shipping delays," "product quality," and "customer service." This output is interpretive—Copilot is making judgment calls about how to group the text. To make it clear what sort of categories you're looking for, you can include examples in your prompt for further guidance. You need to validate it by checking the original comments: do the themes reflect what customers said? Is anything missing or overgeneralized? Copilot gets you there fast; you decide if it's right.

6. Let's practice!

Time to try out performing multi-step analysis through natural language with Copilot!

Create Your Free Account

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.