How to Use Copilot in Excel: Instructions, Features, and Use Cases

Microsoft Copilot has transformed Excel into a more intelligent, collaborative, and productive tool than ever before. In 2025, Copilot is available across the web, desktop, and even mobile Excel apps, integrating natural language AI directly into spreadsheet workflows.

With new features like "Create with Copilot," advanced data cleaning, and even Python analysis built right into the chat pane, using Excel has never been more accessible and powerful.

Who Can Use Copilot in Excel?

In 2025, Microsoft Copilot for Excel is no longer limited to just a handful of business or enterprise subscribers, but instead has expanded its reach dramatically, now being accessible to a much broader spectrum of users across various Microsoft 365 subscription plans and devices. Individuals who subscribe to Microsoft 365 Personal or Family plans are now able to access Copilot, though it is important to note that there is a monthly cap of 60 Copilot requests per user in these plans, which is generally sufficient for home or light professional use but may feel limiting for heavy users or analysts who run hundreds of queries and transformations each month.

Those who require unrestricted or more advanced usage, such as professionals, consultants, or power users, can opt for the Copilot Pro add-on, which can be purchased separately and grants access to Copilot across all major Office applications—including Excel, Word, PowerPoint, and Outlook—removing most artificial usage limits and unlocking the full suite of AI capabilities.

For organizations, Copilot is now a core feature within most Microsoft 365 Business and Enterprise licenses, including Business Basic, Business Standard, and Business Premium, as well as Office 365 E1, E3, and E5, and the parallel Microsoft 365 E1, E3, and E5 plans. The availability also extends to government cloud customers (GCC Moderate and GCC High) and most education tenants (A3/A5), making Copilot a universal tool for enterprise productivity and classroom use. Copilot is fully supported on the latest versions of Excel for Windows (version 2406 or later), Excel for Mac (version 16.84 or later), as well as the always-up-to-date Excel for the web (browser-based), and is steadily rolling out with increased feature parity on mobile Excel apps for both iOS and Android.

If you are unable to find the Copilot feature within your Excel installation, it is advisable to first verify your subscription tier and ensure that your app is fully updated to the latest version; if the Copilot button is still missing, you may need to reach out to your IT administrator or Microsoft 365 account manager to confirm whether your organization has enabled Copilot, as it is possible for administrators to selectively enable or disable access for compliance reasons.

Where to Find Copilot in Excel

Locating Copilot in Excel has become straightforward and user-friendly, thanks to a series of UI enhancements aimed at making the AI both accessible and non-intrusive to traditional workflows. In both desktop and web versions of Excel, you will almost always find the main Copilot button in the upper-right corner of the Excel window.

Clicking this icon launches the Copilot sidebar or chat pane, which occupies the right side of your Excel workspace, providing a conversational interface for typing in prompts, reviewing suggestions, or browsing Copilot’s context-aware sample queries. As of recent updates, Microsoft has also introduced an inline sparkle badge that appears contextually next to your active cell range or selection whenever you highlight a range of data. This inline badge allows users to invoke Copilot directly for specific data areas, ensuring the AI focuses immediately on the content you intend to analyze or transform.

You can always toggle these features on or off according to your preferences by navigating to File > Options > Copilot within Excel, where you will find settings to manage both the persistent sidebar button and the floating inline badge, as well as control over Copilot’s in-app suggestions and preview behaviors. This flexibility ensures that Copilot’s presence never feels disruptive to your normal spreadsheet tasks, while also remaining available at a moment’s notice whenever advanced analysis or automation is required.

What Can You Do With Copilot?

1. Analyze and Summarize Data

Copilot’s analytical capabilities have grown increasingly sophisticated, allowing it to process massive datasets—sometimes containing tens of thousands of rows or hundreds of columns—and instantly surface actionable insights, summaries, and potential errors. For example, if you have imported sales data from several regional offices or multiple fiscal quarters, you can simply highlight your data and prompt Copilot with a request like, “Summarize sales performance by region for Q1 and list the top three outliers in terms of total sales or percentage variance.”

Copilot will then scan the full dataset, recognize and aggregate data by your specified dimension (such as region or quarter), perform statistical analysis to detect which entries deviate significantly from the average, and return a concise, readable summary—often in the form of plain-language text accompanied by supporting mini-tables, bullet points, or even embedded charts. You can further ask Copilot to explain why certain entries are considered outliers, or to break down results by additional filters, such as salesperson, product line, or customer segment. If you notice anomalies or errors, you can immediately prompt Copilot to drill deeper, suggest explanations, or recommend data cleaning actions.

2. Visualize With Charts and Graphs

Rather than spending time navigating through the “Insert Chart” menus or experimenting with different chart types, Copilot can automatically select, generate, and insert the most suitable visualizations based on the structure and nature of your data as well as your descriptive request. For instance, a prompt like “Make a clustered bar chart showing revenue by product line for the past four quarters” instructs Copilot not only to select the correct chart type but also to dynamically identify the relevant time frames and product categories from your dataset.

Copilot will generate the chart directly within your worksheet, using intelligent defaults for labels, legends, and color schemes, while also allowing you to tweak the output by following up with refinements such as, “Switch to a stacked column chart,” “Change the color palette,” or “Include only the top five products.” Additionally, Copilot can suggest chart types if you’re unsure which best fits your analysis; for instance, if you simply ask to “visualize trends,” it may propose a line chart for time-series data or a scatter plot for correlation analysis. The AI’s ability to anticipate your needs and make practical suggestions is especially valuable for users who are less familiar with Excel’s advanced charting tools.

3. Generate or Explain Formulas

One of Copilot’s most widely appreciated features is its ability to write, explain, and even debug formulas in Excel, dramatically reducing the learning curve for users who struggle with complex formula syntax or multi-step calculations. If you describe your objective in plain language—such as, “Write a formula that calculates year-over-year growth in column C based on values in columns B and A,”—Copilot will parse your request, identify the correct cells and ranges, and produce the exact formula (e.g., =(C2-B2)/B2 for growth rate).

Furthermore, Copilot can provide step-by-step explanations for both standard and advanced formulas, breaking down each function or operator in easy-to-understand terms. For instance, if you enter a prompt like, *“What does this formula do: =SUMPRODUCT((A2:A100>0)B2:B100)?” Copilot will not only describe its overall effect (such as summing the values in column B wherever the corresponding entry in column A is positive), but also clarify the purpose of each component and how the logical test operates within the SUMPRODUCT array context. This makes Copilot an invaluable tutor for both beginners seeking to learn Excel formulas and experienced analysts working to troubleshoot or document complex calculations.

4. Clean and Transform Data

Data imported into Excel from other systems—whether it’s a CRM export, ERP report, or CSV download—often requires significant cleaning and reformatting before meaningful analysis can begin. Copilot’s Clean Data feature has become a game-changer for these scenarios, as it can automate many of the most tedious, error-prone data preparation tasks. Suppose you prompt Copilot with, “Remove duplicates from column F and standardize date formats in columns D and E,” the AI will scan for duplicate entries (based on your specified column), eliminate redundant rows, and reformat dates across multiple columns to a consistent format, such as ISO (YYYY-MM-DD) or locale-appropriate short/long forms.

Copilot always provides a preview of the proposed changes, highlighting which rows or cells will be affected, allowing you to review, accept, modify, or undo the transformations before they are finalized. This preview step is essential for data integrity and ensures that large batch edits never result in accidental loss or corruption of important information. For more advanced cleaning, Copilot can also handle tasks such as text normalization, missing value imputation, outlier detection, blank row removal, and out-of-range value flagging. For instance, you might ask Copilot to, “Highlight values above the 95th percentile in this range and flag entries with missing product codes,” receiving both a color-coded worksheet and a summary of findings.

5. Build Templates and Scenarios

The introduction of the Create with Copilot function means that, instead of building complex templates or scenario models from scratch, users can now describe the end goal in natural language and let Copilot construct a structured, formula-driven worksheet tailored to their requirements. If you enter a prompt such as, “Create a SaaS revenue forecast template with separate sections for MRR, churn, and expansion revenue,” Copilot will assemble a multi-section sheet with labeled headers, pre-populated formulas, and even notes or documentation cells explaining how to use or extend the template.

For scenario analysis, Copilot can set up what-if models or sensitivity tables based on user parameters, such as, “Set up a what-if analysis to model a 15% cost increase in raw materials for FY26,” generating the appropriate input cells, linked formulas, and data tables to visualize the impact of different cost assumptions on bottom-line results. The AI not only builds the required structures but can suggest additional fields or metrics you might want to include, ensuring templates are comprehensive, professional, and immediately ready for further customization.

6. Forecasting and Advanced Analytics

With continued integration of machine learning algorithms and advanced statistical engines, Copilot in Excel is now capable of performing in-grid forecasting and time-series analysis directly within the workbook, using Excel’s powerful Forecast/ETS functions. When prompted, “Forecast next quarter’s operating expenses based on this historical data,” Copilot will automatically detect the relevant time series, apply appropriate smoothing or seasonal adjustments, and insert forecasted values in a new column or range, often complete with confidence intervals or trendline visualizations.

For users seeking deeper analysis, Copilot can conduct regression, clustering, or outlier detection, and—if your Excel installation has Python enabled—even execute full-featured statistical scripts from within the chat pane. For example, you can prompt, “Run a linear regression of Sales on Marketing Spend and plot the results,” and Copilot will use Python’s pandas and statsmodels libraries to perform the regression, display coefficients, calculate R-squared and p-values, and generate a publication-quality scatterplot with best-fit line.

When analytical requirements exceed the scope of Excel—such as when you need to cross-analyze data from multiple, linked workbooks, or build complex, multi-variable predictive models—Copilot will recommend transitioning your analysis to Power BI, where Copilot for Power BI unlocks even more advanced AI features, including natural language querying across datasets, multi-table forecasting, and AI-driven anomaly detection.

7. Python in Copilot (Advanced Users)

For organizations that enable the Python in Excel integration, Copilot can become an even more powerful analytics engine, allowing advanced users to perform tasks that would otherwise require exporting data to a separate Python environment. This means that, within the familiar Excel interface, users can now leverage Copilot to both write and execute Python code for a wide array of statistical, financial, and data visualization tasks.

You might request, “Calculate the correlation matrix for columns B through G using Python,” and Copilot will automatically generate the necessary pandas DataFrame code, run the computation securely within Excel’s sandboxed Python runtime, and return the resulting correlation table as a formatted range or object in your sheet. Similarly, a prompt like, “Plot a histogram of customer ages and overlay a normal distribution curve,” will yield a matplotlib or seaborn chart embedded directly in your workbook, complete with dynamic links to your source data. This hybrid approach eliminates the need for context-switching between applications and brings advanced, scriptable analytics directly into everyday Excel workflows.

Language and Accessibility

In its latest incarnation, Copilot has vastly broadened its support for different languages and accessibility needs, ensuring that users from around the world and across all backgrounds can take full advantage of its AI-powered features. The Copilot interface and underlying prompt engine now support over 40 languages, including widely spoken ones such as Spanish, French, German, Italian, Japanese, Chinese, Arabic, and Portuguese, as well as regional and less common languages like Welsh, Basque, or Thai.

Not only can users interact with Copilot in their preferred language, but the AI also adapts to local number formats, date conventions, and even localized data validation rules, making it practical for multinational teams working on joint projects or for organizations with diverse user bases. Accessibility enhancements include high-contrast visual themes, improved keyboard navigation for those who rely on assistive technologies, and compatibility with screen readers, ensuring that Copilot’s suggestions, previews, and results are available to users with visual impairments or motor difficulties.

Privacy and Security Considerations

Microsoft has reinforced Copilot’s privacy and compliance framework, ensuring that all data processed by Copilot in Excel—regardless of whether you are using a Personal, Family, Business, or Enterprise plan—remains within the secure Microsoft 365 “commercial boundary.” This means your spreadsheet data is never sent to external servers for training Microsoft’s AI models, nor is it shared with third parties or used to improve the foundational models behind Copilot.

Organizational administrators retain full control over Copilot deployment, with tools for auditing usage, enabling or disabling Copilot by user or group, and enforcing compliance with industry regulations or local data residency laws. For highly regulated sectors, Microsoft provides granular controls for data access, audit logs for Copilot interactions, and assurance that any Copilot request is logged, traceable, and can be reviewed for compliance. End users are always reminded to avoid uploading confidential, proprietary, or sensitive data unless their Microsoft 365 environment is appropriately secured, and the Copilot Trust Center provides ongoing updates regarding AI privacy practices and compliance certifications.

Tips for Getting the Best Results

To extract maximum value from Copilot in Excel, consider adopting a few best practices that are particularly important as the platform becomes more sophisticated and versatile. Be specific and descriptive in your prompts; clearly identify the columns, ranges, sheets, or time periods you want Copilot to focus on, as this leads to more accurate and context-aware results. Don’t hesitate to refine your requests through iterative follow-ups—the Copilot chat interface is designed for conversational, multi-turn exchanges, meaning you can clarify, expand, or correct your query without starting over from scratch.

Experiment with exploratory prompts—for example, instead of specifying exactly what analysis you want, try asking Copilot to “highlight trends,” “find anomalies,” or “suggest important metrics” based on your dataset. When working with Copilot’s data cleaning or transformation features, always make use of the preview and undo options to review the suggested changes before committing, reducing the risk of accidental edits or lost information. For power users, leveraging the Python integration opens up advanced analytics previously unavailable in Excel, so consider learning a few common Python commands or workflows to unlock the full potential of Copilot’s hybrid capabilities.

Limitations and Known Issues

Despite its many strengths, Copilot in Excel is not without certain limitations and caveats. Users of Personal and Family plans must be aware of the monthly cap of 60 Copilot requests per user, which, while generous for occasional or light use, may not suffice for individuals working on large projects, extensive modeling, or continuous analysis. Copilot is designed to operate exclusively within the boundaries of the workbook’s data and cannot natively access or integrate with external databases, APIs, or real-time web data feeds (for those use cases, Microsoft recommends using Power Query or Power BI).

When handling extremely large or complex workbooks—those with hundreds of thousands of rows, dozens of linked worksheets, or dense arrays of formulas—Copilot’s performance may degrade, and some operations could become slower or produce less accurate results, as the AI is forced to sample or summarize the underlying data. While the rollout of advanced features such as Python scripting, multi-turn memory, and template creation is ongoing, not all features are available in all environments or regions at once; your experience may vary depending on admin settings, beta program participation, or regional compliance restrictions.

What’s New in 2025

2025 has brought several transformative updates to Copilot in Excel, making the platform far more versatile and intuitive than ever before. The Create with Copilot feature allows users to instantly generate structured templates and scenario models by describing their objectives, with the AI constructing fully-formatted, formula-driven tables and dashboards complete with documentation cells and suggested refinements. The new Clean Data workflows automate not only duplicate removal and date standardization, but also advanced normalization, outlier detection, and missing value handling, making data preparation both faster and more reliable.

Python in Copilot further blurs the line between spreadsheet analytics and advanced data science, empowering users to run sophisticated statistical tests, regressions, and visualizations without ever leaving Excel. The addition of the Inline Sparkle Badge makes Copilot available exactly where you need it—right next to any selected range or cell—streamlining contextual analysis and reducing the need to jump between menus or panes. Lastly, expanded language support ensures that Copilot’s capabilities are available to users in more countries and industries than ever before, cementing its role as a global productivity booster.

Example: Step-by-Step Use Case

To see how Copilot can streamline and enhance your workflow in practice, imagine you’re tasked with analyzing a complex sales dataset spanning multiple regions and time periods. First, open Excel on your desktop or in your browser, ensuring you’re logged into a supported Microsoft 365 account with Copilot enabled. Once your data is loaded, click the Copilot icon in the upper-right corner to launch the sidebar, or select your primary table and tap the inline sparkle badge to focus Copilot’s attention on the active range.

Next, type a prompt such as, “Summarize the top trends in this dataset and create a bar chart for the five regions with the highest sales over the last three quarters.” As soon as you press Enter, Copilot will analyze your selected data range or table, automatically grouping and aggregating the sales figures by region and time period. It will then generate a detailed summary in natural language, listing the regions that have demonstrated the highest sales, describing the key growth or decline patterns observed quarter over quarter, and pointing out any notable exceptions or anomalies, such as unexpected drops or surges.

In addition to the written summary, Copilot will insert a professionally formatted bar chart directly into your worksheet, automatically labeling the axes, applying distinct colors for each region, and selecting an appropriate title. If you would like to customize the output further, you can immediately follow up with another prompt, for example, “Change the bar chart to a stacked column chart and add data labels for each bar,” or, “Filter the chart to show only regions with over $1 million in sales.” Copilot’s multi-turn chat memory means you don’t have to repeat the context or range each time—it remembers your previous request and works seamlessly with your workflow.

Suppose, while reviewing the summary, you notice some unusual figures or suspect there may be outliers in the sales data for certain months. You can then type, “Can you highlight any outlier months where sales were more than two standard deviations above or below the quarterly average?” Copilot will run the statistical analysis, visually flag the cells that meet your outlier criteria, and explain in the chat pane which months are considered statistical anomalies and why.

If you realize that some of your raw data may contain inconsistencies or formatting issues—perhaps due to manual entry errors or system exports—you can further instruct Copilot, “Clean this table by removing duplicate entries in the customer ID column, standardize all dates in column D to the YYYY-MM-DD format, and flag any missing values in the product category field.” Copilot will present you with a preview of all the proposed changes, let you accept or undo them in one step, and then document the edits made, providing a summary of the rows or columns affected so you maintain full transparency over your data transformation process.

For those with access to advanced features, if your organization has enabled Python in Excel, you could leverage even deeper analytics. For example, after cleaning your dataset, you might say, “Using Python, calculate a correlation matrix for all numerical fields in this table, and highlight any relationships with a correlation coefficient above 0.7.” Copilot will write the Python code for you, execute it securely within Excel, and present the resulting matrix in a clear, formatted range—making it simple to spot strong statistical relationships between variables, such as the link between marketing spend and sales growth.

As you move forward, perhaps preparing for a management presentation, you might direct Copilot to “Build a scenario analysis table showing how a 10%, 20%, and 30% increase in regional marketing budgets would impact projected Q4 sales,” and the AI will set up a sensitivity table, populate it with dynamic formulas, and even generate scenario-based charts. You can iterate on these prompts, refine your assumptions, and generate documentation or explanatory notes—all within the same conversational flow.

Throughout this process, Copilot ensures that every action it proposes can be reviewed, modified, or reversed, empowering you with both speed and control, and letting you focus more on strategic analysis and less on manual formula construction or data wrangling.

________

FOLLOW US FOR MORE.

DATA STUDIOS

datastudios.org