📚 What you'll do in this training

Four story-driven scenarios, each scoped to a real Finance task you'd be asked to do at a clothing retailer. Each scenario is a sequence of guided labs — you don't just learn a feature, you finish a deliverable.

Scenario A — Close the Month at Zava (Basic)

  1. A1 — Clean a messy GL extract
  2. A2 — Build the variance vs. budget view
  3. A3 — Age the AR and prioritize collections
  4. A4 — Generate a one-page CFO summary

Scenario B — The Board Meeting in 10 Days (Intermediate)

  1. B1 — Decompose revenue into a price/volume/FX bridge
  2. B2 — Extend the cash forecast and flag breach weeks
  3. B3 — Score customer health and retention
  4. B4 — Compose the board KPI page

Scenario C — Forensic Spend Review (Advanced — uses COPILOT() function)

  1. C1 — Find duplicate invoices
  2. C2 — Vendor concentration (Pareto)
  3. C3 — Payment timing anomalies
  4. C4 — Findings memo for the Controller

Scenario D — Cowork Capstone (One prompt, four deliverables)

  1. D1 — Q2 FY26 variance read & close-risk brief
💡 How to use this guide: Each lab is collapsible — click the title to expand. Every prompt has a Copy button so you can paste it straight into Excel Copilot. Work the scenarios in order — they build on each other, going from a single prompt to a full multi-deliverable Cowork run.
Quick Start — Files, prerequisites, and ground rules
⏱️ 5 minutes 🎯 Setup & orientation 🛠️ Read this before Scenario A
📂 Workbooks you'll need: Save all four to your OneDrive or SharePoint before starting. Copilot does not work with files saved only on your local drive.

About Zava: Zava is a fictional global clothing retailer — apparel, footwear, accessories, and outerwear sold through company-owned stores, ecommerce, and a wholesale channel. The labs use Zava's GL, budget, AR, AP, and customer data so the work feels like the real month-end at a CPG / retail business.

Before you start — check these

  • You have a Microsoft 365 Copilot license
  • You're using Excel for the web, Windows, Mac, or iPad (M365 version)
  • Files saved to OneDrive or SharePoint
  • Internet connectivity
  • Data formatted as an Excel table (Ctrl+T) where possible
💡 TIP: Convert ranges to tables with Ctrl+T before prompting. Copilot understands column names and boundaries much better when data is in a table.
🔴 IMPORTANT: Scenario C uses the new COPILOT() function, currently in Preview. You'll need a Microsoft 365 Copilot license and the right update channel (Beta / Current Channel Preview). If you get #NAME?, the function isn't enabled on your Excel build yet — you can still do A, B, and D.
⚠️ WARNING — verify everything: Copilot output is a productivity tool, not a substitute for professional judgment. Sanity-check every number it quotes before you act on it. Use Ctrl+Z the moment a result looks wrong.

Scenario A — Close the Month at Zava

You're a Senior Financial Analyst at Zava. It's the 4th business day of May. Your Controller wants the variance pack, AR aging, and a CFO summary by EOD.
🎯 Level: Basic ⏱️ ~45 minutes 📄 Workbook: Zava_Scenario_A_MonthClose.xlsx 👤 Senior Financial Analyst
The story. It's the 4th business day of May 2026. Your Controller drops a Teams message: "I need the May variance read, an AR aging with the top 10 collections priorities, and a one-page summary for the CFO — by end of day." The ERP just spit out the GL, but the department names are inconsistent and a few rows are blank. Normally this is a full day of work. Today you're going to do it in about 45 minutes with Copilot.
Lab A1 — Clean a messy GL extract
⏱️ 10 minutes 🎯 Goal: Standardize departments & fill blanks 📄 Sheet: GL_Actuals

Open Zava_Scenario_A_MonthClose.xlsx and go to the GL_Actuals sheet. Skim it — you'll see "Marketing", "Mktg", "MARKETING", "Marketing Dept" all in the same column, and a few blank cells in Department and Region. Copilot will fix it.

Exercise A1.1 — Find the inconsistencies

💬 Try this prompt
List every unique value in the Department column and group together the ones that look like the same department spelled differently.

Exercise A1.2 — Standardize department names

💬 Try this prompt
Add a column called "Department (Clean)" that maps every Department value to a single canonical name. Treat Mktg, MARKETING, and Marketing Dept as Marketing. Treat E-commerce, eCommerce, and Ecomm as Ecommerce. Treat Store Operations, Store Ops, and STORES as Stores. Trim trailing spaces.

Exercise A1.3 — Flag the blanks

💬 Try this prompt
Highlight rows where Department or Region is blank in yellow so I can chase the GL team for the coding.
💡 TIP: When you ask Copilot for a "clean" column, look at the formula it writes — usually SWITCH or nested IF chains. You can copy that pattern next time you get a similar feed.
Lab A2 — Build the variance vs. budget view
⏱️ 12 minutes 🎯 Goal: Actual vs Budget by department, with variance $ and % 📄 Sheets: GL_Actuals, Budget

You have May actuals on GL_Actuals and the approved budget on Budget. You need a clean Actual vs Budget view by department, with variance dollars and variance percent.

Exercise A2.1 — Pivot actuals by department

💬 Try this prompt
Using the cleaned Department column on GL_Actuals, create a summary table that totals Amount by Department for May 2026.

Exercise A2.2 — Join in the budget

💬 Try this prompt
For each department in my summary, add a column with the total Budget Amount from the Budget sheet for May 2026. Then add Variance ($) = Actual - Budget, and Variance (%) = Variance / Budget.

Exercise A2.3 — Highlight the big movers

💬 Try this prompt
Highlight rows where absolute Variance % is greater than 10% in red, and rows where it's between 5% and 10% in yellow.

Exercise A2.4 — Chart it

💬 Try this prompt
Create a horizontal bar chart of Variance ($) by Department, sorted from most unfavorable to most favorable.
⚠️ WARNING: Sign convention matters. Revenue is typically posted as a credit (negative) in the GL. If your variance signs look flipped, ask Copilot: "How are revenue accounts signed in this data? Adjust the variance so favorable revenue variances are positive."
Lab A3 — Age the AR and prioritize collections
⏱️ 10 minutes 🎯 Goal: Aging buckets and a top-10 chase list 📄 Sheet: AR_Invoices

Switch to the AR_Invoices sheet. Treat "today" as May 8, 2026 for all aging math.

Exercise A3.1 — Add aging buckets

💬 Try this prompt
Treating today as May 8, 2026, add two columns: "Days Past Due" (days between Due Date and today, negative if not yet due) and "Aging Bucket" with values Current, 1-30, 31-60, 61-90, 90+.

Exercise A3.2 — Summarize by bucket

💬 Try this prompt
Create a table showing total Balance Due by Aging Bucket. Add a percent-of-total column.

Exercise A3.3 — Build the chase list

💬 Try this prompt
List the top 10 invoices by Balance Due that are 31 days or more past due. Show Customer, Invoice Number, Days Past Due, and Balance Due.

Exercise A3.4 — Customer concentration

💬 Try this prompt
Which three customers have the largest total past-due balance? Show their total Balance Due and the number of past-due invoices each.
💡 TIP: When asking for "top N" results, always specify the sort field. "Top 10 invoices" is ambiguous — "top 10 by Balance Due" is not.
Lab A4 — Generate a one-page CFO summary
⏱️ 10 minutes 🎯 Goal: Narrative commentary the CFO can paste into a deck 📄 All sheets

You've got the cleaned GL, the variance view, and the aging. Time to write the commentary.

Exercise A4.1 — A single executive paragraph

💬 Try this prompt
Read the variance summary by department and write a one-paragraph executive summary for the CFO. Lead with the biggest unfavorable variance, then the biggest favorable, then anything else worth flagging. Keep it under 120 words.

Exercise A4.2 — Structured commentary

💬 Try this prompt
Write the CFO summary in four short sections: Revenue, Gross Margin, Operating Expenses, AR Risk. One short paragraph per section. Reference specific dollar amounts and percentages from the variance table.

Exercise A4.3 — Ask the obvious follow-up

💬 Try this prompt
Based on the variances and AR aging, what are the three questions the CFO is most likely to ask me, and what data should I have ready to answer them?
⚠️ WARNING: Copilot will sometimes make up a driver that sounds plausible but isn't in the data ("higher freight costs due to tariffs"). Strike anything you can't trace back to a row in the workbook before you send the summary to the CFO.

Scenario B — The Board Meeting in 10 Days

You're VP of Finance. The CFO wants a revenue bridge, a 13-week cash forecast, customer health metrics, and a one-page board KPI summary. You have 10 days and one workbook.
🎯 Level: Intermediate ⏱️ ~60 minutes 📄 Workbook: Zava_Scenario_B_BoardPrep.xlsx 👤 VP of Finance
The story. The Zava board meets in 10 days. The CFO sends a one-line ask: "Bridge, cash, customers, KPIs — board ready." You have actuals through Q2 FY26, a 12-week cash file the Treasury team handed off, and a customer revenue cut by quarter. Translate it into four board-ready artifacts.
Lab B1 — Decompose revenue into a price/volume/FX bridge
⏱️ 15 minutes 🎯 Goal: Explain Q2 vs Q1 revenue movement 📄 Sheet: Revenue_Bridge

Open Zava_Scenario_B_BoardPrep.xlsx and go to Revenue_Bridge. You have prior quarter and current quarter revenue, units, average prices, and an FX rate change per product line.

Exercise B1.1 — Compute the three effects

💬 Try this prompt
For each product line, add three columns: Volume Effect = (Current Units - Prior Units) * Prior Avg Price; Price Effect = (Current Avg Price - Prior Avg Price) * Current Units; FX Effect = FX Rate Change * Prior Quarter Revenue. Then add a Total Change column that equals Current Revenue - Prior Revenue and check it ties to the sum of the three effects.

Exercise B1.2 — Validate the math

💬 Try this prompt
For each row, calculate the difference between Total Change and (Volume Effect + Price Effect + FX Effect). Highlight any row where the difference is more than $50,000 in red — that's a sign the decomposition isn't tying out.

Exercise B1.3 — Build the bridge chart

💬 Try this prompt
Build a waterfall chart showing total Prior Quarter Revenue, then a bar for Volume Effect, Price Effect, and FX Effect, ending in Current Quarter Revenue.

Exercise B1.4 — Plain-English narrative

💬 Try this prompt
Write a three-sentence summary explaining the Q2 revenue movement: total dollar change, the dominant driver (volume, price, or FX), and the one product line that is the biggest outlier.
💡 TIP: Price/volume bridges are a classic interview question and a board staple. Once you've seen Copilot write the formula once, you'll recognize the pattern in every revenue analysis you do.
Lab B2 — Extend the cash forecast and flag breach weeks
⏱️ 15 minutes 🎯 Goal: 12 → 13 weeks, threshold flagging, narrative 📄 Sheet: Cash_Forecast

Open the Cash_Forecast sheet. The Treasury team gave you 12 weeks. You need 13. The board minimum cash threshold is $8,000,000.

Exercise B2.1 — Extend the forecast one more week

💬 Try this prompt
Add a 13th forecast week to the table. Use the 12-week averages for Operating Receipts, Operating Disbursements, CapEx, and Debt Service. Beginning Cash should pull from the prior week's ending cash. Recalculate the Ending Cash Balance.

Exercise B2.2 — Threshold flagging

💬 Try this prompt
Add a "Below Threshold" column that says YES whenever Ending Cash Balance is below $8,000,000. Highlight any YES row in red across the whole row.

Exercise B2.3 — The line chart

💬 Try this prompt
Create a line chart of Ending Cash Balance by Week Ending. Add a horizontal reference line at $8,000,000.

Exercise B2.4 — What's driving the dip

💬 Try this prompt
For the week with the lowest ending cash balance, explain in two sentences what drove the dip. Compare its Disbursements and CapEx to the 13-week average.
📝 NOTE: If Copilot doesn't add the threshold reference line on the chart, ask: "Add a constant series at $8,000,000 to overlay the threshold." Plain language usually works.
Lab B3 — Score customer health and retention
⏱️ 15 minutes 🎯 Goal: Retention rate, growth tiers, churn list 📄 Sheet: Customer_Cohorts

Open Customer_Cohorts — a mix of wholesale (B-prefix IDs) and DTC ecommerce (C-prefix IDs) customers with prior and current quarter revenue.

Exercise B3.1 — Net revenue retention

💬 Try this prompt
Calculate Net Revenue Retention = sum of Current Quarter Revenue divided by sum of Prior Quarter Revenue, split by Channel (Wholesale vs DTC Ecommerce). Show both numbers and the ratio as a percent.

Exercise B3.2 — Bucket customers

💬 Try this prompt
Add a column called "Health" that buckets each customer based on Current Quarter Revenue vs Prior Quarter Revenue: "Churned" if current is zero; "Contracting" if current is less than 90% of prior; "Flat" if 90-110%; "Growing" if 110-130%; "Expanding" if more than 130%.

Exercise B3.3 — Health distribution

💬 Try this prompt
Create a stacked bar chart showing the count of customers in each Health bucket, split by Channel.

Exercise B3.4 — Top churn losses

💬 Try this prompt
List the top 5 churned customers (Health = "Churned") by Prior Quarter Revenue. These are our biggest at-risk dollars from last quarter.
💡 TIP: Net Revenue Retention is a single number that captures customer health better than any chart. If it's above 100%, your existing book is growing on its own. Below 100%, you're leaking.
Lab B4 — Compose the board KPI page
⏱️ 15 minutes 🎯 Goal: One-page board summary 📄 All sheets

You have the bridge, the cash forecast, and the customer health metrics. Roll it up.

Exercise B4.1 — The five headline KPIs

💬 Try this prompt
Give me five board-level KPIs I can put on a single slide, using this workbook. For each KPI, show the value, the prior quarter comparison, and whether the trend is favorable. Cover: revenue growth, FX-neutral revenue growth, lowest forecast cash, net revenue retention, customer count change.

Exercise B4.2 — The narrative box

💬 Try this prompt
Write a 150-word board narrative that ties together the revenue story, the cash position, and customer health. Be direct about risks. Reference specific numbers from the workbook.

Exercise B4.3 — Anticipate board questions

💬 Try this prompt
Based on these KPIs and trends, what are the four most likely questions the board will ask, and what is the best one-line answer for each?

Stretch — your own KPI

Pick one KPI you'd add that isn't in the list above. Write the prompt to compute it from the workbook. Compare with the person next to you.

⚠️ WARNING: Board content is high-stakes. Reconcile every figure on your one-pager to a cell in the workbook. If a number can't be traced, remove it.

Scenario C — Forensic Spend Review

Internal Audit flagged Procurement spend as a risk. You have 5 days to find duplicates, identify consolidation opportunities, evaluate payment timing, and write a findings memo for the Controller.
🎯 Level: Advanced ⏱️ ~60 minutes 📄 Workbook: Zava_Scenario_C_SpendReview.xlsx 🧪 Uses COPILOT() function 👤 Senior FP&A / Internal Audit
The story. Internal Audit flagged Procurement as a risk area in the last review. You've been seconded for a 5-day rapid review on 12 months of accounts payable activity. The Controller wants a one-page findings memo — duplicates, vendor concentration, payment timing, and a number for potential savings.
🔴 IMPORTANT: Scenario C uses the COPILOT() function in some exercises. It's in Preview — if you get #NAME?, your channel doesn't have it yet. You can still complete every lab using side-pane prompts instead of formulas.
Lab C1 — Find duplicate invoices
⏱️ 12 minutes 🎯 Goal: Surface same-vendor, same-amount, close-date pairs 📄 Sheet: AP_Transactions

Open Zava_Scenario_C_SpendReview.xlsx and go to AP_Transactions. Roughly 250 transactions over 12 months. There are 5 duplicate pairs hidden in here. Find them.

Exercise C1.1 — Group by vendor + amount

💬 Try this prompt
Find every pair of rows in AP_Transactions where the Vendor Name and Amount are identical and the Invoice Date is within 7 days of each other. List both Invoice Numbers, the Vendor, the Amount, and the date gap.

Exercise C1.2 — Quantify the exposure

💬 Try this prompt
Sum the total dollars of the duplicate pairs you found. If we paid both invoices in each pair, what's our maximum overpayment exposure?

Exercise C1.3 — Highlight in place

💬 Try this prompt
Highlight all duplicate rows in AP_Transactions in yellow so I can quickly find them on the source sheet.
💡 TIP: Real duplicate detection in AP also matches invoices where the amount is identical but the invoice number is slightly different (a transposed digit, a leading zero). Ask Copilot to find "near-duplicates" as a stretch.
Lab C2 — Vendor concentration (Pareto)
⏱️ 12 minutes 🎯 Goal: Identify the 20% of vendors that drive 80% of spend 📄 Sheet: AP_Transactions

Exercise C2.1 — Rank by spend

💬 Try this prompt
Create a table that sums Amount by Vendor Name, sorted from largest to smallest. Add a running cumulative percent of total spend.

Exercise C2.2 — Find the 80/20 line

💬 Try this prompt
How many vendors make up the first 80% of total spend? List them.

Exercise C2.3 — Use the COPILOT() function to categorize

In an empty cell, try the COPILOT() function. First, pull the unique vendor list to a column (say H4:H40), then:

💬 Type this formula
=COPILOT("Categorize each vendor as Strategic (big-spend, mission-critical), Preferred (mid-spend, repeatable services), or Tail (small-spend, infrequent)", H4#)

Exercise C2.4 — Pareto chart

💬 Try this prompt
Create a Pareto chart of total spend by vendor — bars in descending order, with a cumulative percent line overlay.
📝 NOTE: Tail spend is where consolidation savings hide. If 20 vendors each charge $2,000 for office supplies, you're probably overpaying — that's the procurement opportunity.
Lab C3 — Payment timing anomalies
⏱️ 12 minutes 🎯 Goal: Find vendors paid too fast or too slow 📄 Sheet: AP_Transactions

Standard terms at Zava are Net 30. Anything paid in under 15 days is suspiciously fast. Anything paid past 50 days is suspiciously slow.

Exercise C3.1 — Days to pay

💬 Try this prompt
Add a column "Days to Pay" = Payment Date - Invoice Date.

Exercise C3.2 — Flag the outliers

💬 Try this prompt
Add a "Flag" column: "Suspiciously Fast" if Days to Pay is less than 15, "Suspiciously Slow" if Days to Pay is more than 50, otherwise "Normal".

Exercise C3.3 — Which vendors get paid fast

💬 Try this prompt
List the top 10 vendors by count of "Suspiciously Fast" payments. For each, show the average Days to Pay and the total dollar amount paid fast.

Exercise C3.4 — Which vendors get paid slow

💬 Try this prompt
Same analysis for "Suspiciously Slow" — top 10 vendors by count, average Days to Pay, total amount.

Exercise C3.5 — Estimate the discount giveaway

💬 Try this prompt
If we assume a 2% early-payment discount was available on every "Suspiciously Fast" payment but we paid before claiming it, estimate the dollar amount we left on the table.
⚠️ WARNING: Fast payments aren't always a problem (sometimes a vendor has a true 2/10 discount). Slow payments aren't always bad (sometimes the invoice was disputed). Treat the list as questions to ask, not findings.
Lab C4 — Findings memo for the Controller
⏱️ 12 minutes 🎯 Goal: One-page memo with quantified findings 📄 Sheet: AP_Transactions

Exercise C4.1 — Roll up the findings

💬 Try this prompt
Write a one-page findings memo to the Controller covering four sections: Duplicate Invoices (count, dollar exposure), Vendor Concentration (top vendors as % of spend, consolidation opportunity), Payment Timing (count and dollars paid fast / paid slow), and Estimated Savings (a single recovery range we should chase). Keep it under 400 words.

Exercise C4.2 — Three recommended actions

💬 Try this prompt
Recommend three concrete next steps the Controller should take based on these findings. Each one should have an owner, an ETA, and the expected dollar impact.

Exercise C4.3 — Self-critique

💬 Try this prompt
Read your own findings memo back and flag any claim that isn't directly supported by a row or summary in the workbook. Rewrite each flagged sentence so it stays defensible.
💡 TIP: The "self-critique" prompt is one of the most underused moves in Copilot. Having the model audit its own draft catches the hallucinations before your reviewer does.

Scenario D — Cowork Capstone

One prompt. Four deliverables. A full Cowork run that closes a quarter end-to-end.
🎯 Level: Capstone ⏱️ ~25 minutes 📄 Workbook: EDG_Finance_Performance.xlsx 🤝 Cowork 👤 VP Finance / FP&A Director
The story. You've finished the Zava scenarios — now switch hats. In Scenarios A–C you used Copilot as a single-step helper inside Excel. Cowork is different: you give Copilot one rich prompt and it produces multiple connected deliverables (a variance workbook, a CFO brief, a QBR deck, and a one-page HTML dashboard) in a single agentic run. This capstone uses an Edgewell Personal Care dataset to push the model harder than the Zava data does.
Lab D1 — Cowork: one prompt, four deliverables
⏱️ 20–25 minutes 🎯 Goal: Run a multi-artifact Cowork scenario end-to-end 👤 Persona: VP Finance / FP&A Director 💼 Finance use case 🤝 Cowork
📂 File for this lab: EDG_Finance_Performance.xlsx — download this workbook and save it to your OneDrive or SharePoint before starting. It contains Edgewell Personal Care's Q2 FY26 Budget vs Actual data by department × month, key drivers, and an open close-task list.

In Labs 1–5 you used Copilot as a single-step helper inside Excel. Cowork is different — you give Copilot one rich prompt and it produces multiple connected deliverables (a variance workbook, a CFO brief, a QBR deck, and a one-page HTML dashboard) in a single agentic run. The persona below is a VP of Finance closing Q2 FY26 books for Edgewell Personal Care and preparing the variance review for the CFO.

The scenario

Company: Edgewell Personal Care (NYSE: EPC) — ~$2.2B in revenue across three segments (Wet Shave, Sun & Skin Care, Feminine Care) plus DTC brands (Billie, Jack Black). Fiscal year ends September 30.

What's pressuring the quarter:

  • Wet Shave — volume softness in North America private label
  • Sun & Skin Care — residual Banana Boat recall echo holding back NA shelf placement
  • Cuautitlan plant — operator attrition driving unfavorable labor variance
  • A&P spend — running 80 bps over plan to defend share
  • Billie CAC — $28.40 vs $24.00 budgeted
  • Productivity savings — ~380 bps GM tailwind partially offsetting ~195 bps materials inflation

Prompt Title: Q2 FY26 Variance Read & Close-Risk Brief for Edgewell

Exercise D1.1 — The Cowork prompt

Open Copilot Chat (Microsoft 365 Copilot, not just the Excel side pane). Attach EDG_Finance_Performance.xlsx and paste this prompt:

💬 Cowork prompt — paste in full
You are my FP&A partner for Edgewell Personal Care. We are closing Q2 FY26 and I need to walk the CFO through the variance story and the open close risks tomorrow morning.

STEP 1 — ANALYZE
Read EDG_Finance_Performance.xlsx and build a Q2 FY26 variance read:
• Compare Actual vs Budget by department and by month for Q2 FY26 (Jan–Mar 2026 calendar / fiscal Q2).
• Identify the top 5 favorable and top 5 unfavorable variances in absolute dollars AND as a % of plan.
• Call out anything that crosses ±10% vs plan or ±$500K in absolute terms.
• Map variances back to the named drivers in the file (Wet Shave NA volume, Banana Boat recall echo, Cuautitlan attrition, A&P over plan, Billie CAC, productivity savings).
• Flag the three highest-risk items on the open close-task list.

STEP 2 — OUTPUTS (produce all four)
1. EDG_Finance_VarianceDashboard.xlsx — a clean variance workbook with: a summary tab (Actual / Budget / Variance $ / Variance % by department), a drivers tab tying the largest variances to root causes, and a close-risk tab listing the three highest-risk open tasks with owner, ETA, and impact.
2. EDG_Finance_CFO_Brief.docx — a 2-page memo for the CFO: TL;DR up top, Revenue read, Gross Margin read, Operating Expense read, Close-Risk section, and a Recommendation paragraph.
3. EDG_Finance_QBR_Deck.pptx — a 6–8 slide deck for the quarterly business review: title, Q2 headline, segment view (Wet Shave / Sun & Skin Care / Feminine Care / DTC), GM bridge (materials inflation vs productivity), A&P and CAC slide, close-risk slide, ask/recommendation.
4. EDG_Finance_Dashboard.html — a single-page HTML KPI summary the team can open in a browser: top metrics tiles, the segment variance chart, and the three close-risk items.

STEP 3 — TALK TRACK
At the end, give me a 60–120 second executive talk track I can use to open the CFO meeting. Lead with YTD top line, then the Sun & Skin Care NA swing (12–18% below plan), then the GM story (195 bps materials inflation offset by ~380 bps productivity), then the A&P 80 bps over plan, then the three highest-risk close tasks. Keep it crisp.

Rules:
• Only use numbers and drivers that are in the attached file. If something isn't in the data, mark it [ASSUMPTION] in the deliverable.
• Don't invent customer names, contract values, or forecast numbers.
• Where you make a judgment call, say so in one line and explain why.

What Copilot produces

One Cowork run should hand you back four linked files:

  • EDG_Finance_VarianceDashboard.xlsx — summary, drivers, and close-risk tabs
  • EDG_Finance_CFO_Brief.docx — 2-page memo for the CFO
  • EDG_Finance_QBR_Deck.pptx — 6–8 slide quarterly business review deck
  • EDG_Finance_Dashboard.html — one-page browser-friendly KPI summary
🎤 Executive talk track (what "good" sounds like): "YTD we're tracking roughly in line on top line, but Q2 has a Sun & Skin Care NA swing 12–18% below plan — that's the Banana Boat recall echo still holding back shelf placement. Gross margin is the bright spot under the surface: ~195 bps of materials inflation, fully offset by ~380 bps of productivity from Cuautitlan and Sri Lanka. The watch-out is A&P running 80 bps over plan as we defend share, plus Billie CAC at $28.40 vs the $24 we budgeted. On close, three items are live: Cuautitlan labor variance reconciliation, the Banana Boat returns reserve true-up, and the DTC subscription deferred revenue split. I'd like 48 more hours on the first two before we lock the books."
🔴 IMPORTANT — Assumptions vs Facts: Anything Copilot produces that is NOT in EDG_Finance_Performance.xlsx should be marked [ASSUMPTION] in the deliverables. Customer names, future-period forecast figures, and any contract dollars not in the file are out of bounds. Sanity-check every number against the source workbook before sharing with the CFO.

Stretch — optional follow-ups

Once the main run lands, try these as separate Copilot prompts in the same chat:

💬 Stretch prompt 1 — Q3 forecast scenario
Using the Q2 actuals and the named drivers, build a Q3 FY26 forecast scenario with three cases (low / base / high). Assume Sun & Skin Care NA recovers 30% / 50% / 70% of the gap, Cuautitlan attrition stabilizes by mid-Q3, and A&P holds at the current run rate. Show segment revenue, GM%, and Adjusted EBITDA for each case.
💬 Stretch prompt 2 — one-page close-risk memo
Write a one-page close-risk memo to the Controller covering only the three highest-risk open close tasks. For each: what the risk is, the dollar exposure range, the owner, and the recommended path to resolution before we lock the books.
⚠️ WARNING: Cowork output is a starting draft, not a final deliverable. Always reconcile the numbers in the generated workbook against your source ERP data, review the CFO brief for tone and accuracy, and verify chart figures on the deck slides before presenting. The agent is a force multiplier — not the signer-off.
🎓 Wrap-up & next steps

What you just did

  • Scenario A: Closed a month at Zava — cleaned the GL, built the variance pack, aged AR, and wrote the CFO summary.
  • Scenario B: Prepped for the board — built a revenue bridge, extended the cash forecast, scored customer health, and rolled it into a board KPI page.
  • Scenario C: Ran a forensic spend review — found duplicates, mapped vendor concentration, flagged payment-timing anomalies, and wrote a Controller-ready findings memo.
  • Scenario D (Cowork): Used one Cowork prompt to produce a four-artifact close package end-to-end.

Best practices to take with you

  • Format data as a table before working with Copilot (Ctrl+T).
  • Be specific — column names, thresholds, desired output format.
  • Iterate — start simple, verify, then build complexity.
  • Always review the formula Copilot writes before building on it.
  • Use the self-critique prompt before sending anything to leadership.
  • Keep a prompt journal for recurring tasks — you'll reuse 80% of them next month.
  • Use Ctrl+Z the moment something looks off.

Where to go next

  • Try Copilot on your own data — bring a workbook you actually use this month.
  • Explore Python in Excel with Copilot for advanced statistical work.
  • Try Cowork on your own multi-deliverable scenario (variance + memo + deck).
  • Share your best prompts with the team.

Zava Excel Copilot Training — Guided Labs · Last updated 2026-05-14