← Resources Free guide · Copilot prompts

A business analyst’s day, with Copilot.

From a messy export first thing to the two-minute summary for the executive at the end of the day. Every prompt tells Copilot who to be, what the data looks like, and how to check its own work.

The day

Eight moments. Nine prompts.

Copy a prompt, press Open Copilot to start Microsoft 365 in a new tab, open your workbook and paste it in. Words in [brackets] are yours to fill in.

  1. Copilot in Excel8:00Microsoft Excel

    Get the export ready for analysis

    Data exported from another system rarely arrives clean. Fix the duplicates, dates and spaces first, on a new sheet, so the original stays untouched.

    Act as an Excel data analyst. Clean the table in this sheet so it’s ready for analysis. The data was exported from another system and may contain duplicates, inconsistent date formats, extra spaces and blank rows. Don’t delete any columns or overwrite the original - add cleaned data to a new sheet. Return a cleaned table plus a short list of every change made. Before finishing, check no valid rows were removed, all dates use DD/MM/YYYY and totals match the original.

  2. Copilot in Excel9:00Microsoft Excel

    Build the formulas

    Describe the columns and the result you want, and Copilot writes the formula, fills it down and tests it, so you don’t have to remember the syntax.

    Flag overdue invoices

    Act as an Excel formula expert. Create a formula in a new column that flags any invoice overdue by more than 30 days. Column A holds invoice dates, column B the amount and column C payment status (Paid or Unpaid). Use functions available in Excel 365, avoid helper columns and keep it readable. Add the formula, fill it down, and explain in two sentences how it works. Before finishing, test it against at least three rows, including a paid invoice and a blank date.

    Match records with XLOOKUP

    Act as an Excel specialist. Pull customer email addresses from the Contacts sheet into the Orders sheet using the customer ID they share. Some IDs may be missing or contain extra spaces. Use XLOOKUP, trim the IDs, and show “Not found” instead of an error. Add the results as a new column in Orders and list how many IDs didn’t match. Before finishing, spot-check five rows against the Contacts sheet and confirm no #N/A errors remain.

  3. Copilot in Excel10:30Microsoft Excel

    Find what the data is saying

    Before anyone sees a chart, know the trends, the outliers and the cells that need a second look.

    Analyse the data in [range/table name]. Give me: 1. A summary of key trends and outliers 2. Any anomalies or data quality issues, blanks, duplicates, inconsistent formats. 3. Three actionable insights based on the patterns you find. 4. A suggested chart type to visualise the main trend, and why. Assume I’m [your role - e.g. a marketing manager reviewing monthly campaign spend] and keep the summary under 150 words. Flag anything that needs a second look before I present it.

  4. Copilot in Excel11:30Microsoft Excel

    Check spend against budget

    Give Copilot the columns, the threshold and the format you want, and the overspends show up in red with a short summary on top.

    Budget against actual

    Act as a finance analyst. Compare budget against actual spend in this sheet and highlight where we’re off track. Columns show cost category, monthly budget and actual spend. Calculate variance in both £ and %, flag anything more than 10% over budget in red, and keep commentary brief. Add variance columns, apply the formatting and write a 3-line summary of the biggest overspends. Before finishing, check variance signs are correct and no category has been missed.

  5. Copilot in Excel13:30Microsoft Excel

    Forecast the next quarter

    Ask for trend and seasonality only, so the forecast comes from your numbers and not from guesses about the outside world.

    You’re a planning analyst. Forecast the next 3 months of revenue from the monthly figures in this sheet, which cover 24 months with seasonal peaks in November and December. Base it on trend and seasonality only and don’t invent outside factors. Add a forecast table, a line chart of actuals versus forecast and a short note on assumptions. Before finishing, check the forecast follows past seasonal patterns and flag any month that looks unrealistic.

  6. Copilot in Excel15:00Microsoft Excel

    Write the weekly sales update

    A busy manager wants three headlines and what to do about them, not every row of the sales data.

    You’re a commercial analyst reporting to a busy manager. Summarise the key trends in this sales data for a weekly update. Columns cover date, region, product and revenue. Focus on the last 4 weeks versus the previous 4, use plain English and skip jargon. Give me 3 headline insights, a pivot table of revenue by region and product, and one chart showing the weekly trend. Before finalising, check every figure ties back to the source data and each insight suggests an action.

  7. Copilot in PowerPoint16:00Microsoft PowerPoint

    Turn the numbers into slides

    Type / in Copilot to pick the spreadsheet, then ask for a short deck that tells the story: one insight, one chart and one headline per slide.

    Act as a data storyteller. Create 4 slides from /[spreadsheet name] showing our [monthly sales] performance for [team meeting]. The audience wants the headline trends, not every number, so pick the 3 most important insights. Use one chart per slide, a headline that states the insight, and no more than 2 supporting bullets. Add a final slide with recommended actions. Before finishing, check every figure matches the source and each chart is labelled and easy to read.

  8. Copilot Chat16:45Microsoft Copilot

    Brief the executive

    The last job of the day: a long document turned into something an executive can read in two minutes, with the one action they need to take.

    Two-minute executive summary

    Act as a senior business analyst. Summarise this document for a busy executive who has two minutes to read it. Focus on the key decisions, risks and deadlines, ignoring background detail. Use plain UK English, no jargon, under 200 words. Deliver a one-line headline, five bullet points of key takeaways, and a final line stating the single action needed. Before finalising, check every point is drawn from the document, nothing is invented, and the action is clear.

Copilot in Excel works on files saved to OneDrive or SharePoint. Copilot in Excel and PowerPoint depend on your Microsoft 365 Copilot licence and the features your IT team has switched on. Check the figures before anything goes to a manager or the board: you own what you send.

9prompts that check their own workEvery prompt ends with a “before finishing, check…” step
  • Clean data firstDuplicates, dates and spaces are fixed on a new sheet, with a list of every change, before any analysis starts.
  • Answers, not just numbersTrends, variances and forecasts arrive with a short summary and the action each one suggests.
  • Ready for the meetingCharts, slides and a two-minute summary, built from the same source and checked against it.
Before you start

Three habits that make every prompt better.

  • Give Copilot a roleStart with “Act as…” so the answer comes from the right point of view: a formula expert, a finance analyst, a data storyteller.
  • Describe the dataName the columns, the sheets and what might be wrong with them. Copilot only knows what you tell it about the file.
  • Make it check its own workEnd with “Before finishing, check…”: rows kept, totals matching, no errors left. Then spot-check the big numbers yourself.
Make it a team habit

One day is the start, not the finish.

Get your first win with the Quick Win Checklist, then turn it into a habit across your team with The Copilot Playbook.

Free, no-obligation AI readiness call

Want prompts like these for your own team?

We’ll look at how your people actually spend their day and build the prompts and habits that fit it.

Henrico from the Solid Systems team, smiling in a Solid polo shirt