Microsoft Excel is one of the areas where GPT-5.6 can deliver immediate value to business users.
OpenAI states that GPT-5.6 produces more polished and accurate spreadsheets, handles equations and financial models with greater precision, and makes better use of worksheet layout, spacing, hierarchy, and formatting. Microsoft has also selected GPT-5.6 as the preferred model in Microsoft 365 Copilot, noting that it can support deeper Excel analysis while using tokens more efficiently.
The improvement is not limited to generating formulas. GPT-5.6 can help users understand existing workbooks, identify trends, investigate errors, create financial models, prepare management reports, and turn raw business data into actionable insights.
ChatGPT for Excel works through a sidebar inside Excel. It can work with large, multi-sheet workbooks containing formulas, references, assumptions, charts, and formatting. When used with Codex in the ChatGPT desktop app, it can inspect and update sheets, formulas, values, formatting, and charts directly in an open workbook.
Here are some of the most useful Excel applications for businesses.
1. Monthly Sales Performance Analysis
A company may have an Excel file containing thousands of sales transactions with columns such as:
- Transaction date
- Customer name
- Product or service
- Sales representative
- Country or region
- Quantity
- Revenue
- Cost
- Gross profit
Instead of manually creating formulas and PivotTables, users can ask GPT-5.6 to analyse the workbook and identify the most important findings.
For example, GPT-5.6 could help calculate:
- Total monthly revenue
- Month-over-month growth
- Year-over-year growth
- Revenue by product
- Revenue by salesperson
- Average order value
- Gross profit margin
- New versus returning customer revenue
- Top-performing and underperforming regions
- Customers with declining purchases
Example prompt:
Analyse the Sales Data worksheet. Create a monthly sales summary showing revenue, gross profit, gross margin, number of transactions and average order value. Compare the latest month with the previous month and the same month last year. Highlight the five biggest positive and negative changes.
GPT-5.6 can also prepare a management summary explaining the results in plain English.
For example:
Revenue increased by 8.4% compared with the previous month, mainly due to stronger Microsoft 365 sales. However, gross margin declined because more orders were sold under promotional pricing.
This allows management to understand not only what changed, but also the likely business drivers behind the change.
2. Budget Versus Actual Reporting
Finance teams frequently compare actual income and expenses against an approved budget.
A typical workbook may include:
- Department
- Account category
- Monthly budget
- Actual expenditure
- Variance
- Variance percentage
- Forecast
- Management comments
GPT-5.6 can help create the necessary formulas, classify variances and identify areas that require management attention.
Example prompt:
Compare the Budget and Actual columns for every department. Add the dollar variance and percentage variance. Highlight expenses that exceed budget by more than 10%. Create a summary showing the five largest unfavourable variances and explain which departments contributed most to the overspending.
It could also generate a full-year forecast using current spending patterns.
Example prompt:
Based on actual expenditure from January to June, forecast the full-year expenses for each department. Use the existing budget as a reference and clearly state the assumptions used.
This is particularly useful when preparing monthly management accounts or board reports.
3. Cash-Flow Forecasting
Cash-flow planning is critical for small and medium-sized businesses.
GPT-5.6 can help create a cash-flow model that combines:
- Opening bank balance
- Expected customer payments
- Recurring revenue
- Supplier payments
- Payroll
- Rental
- Taxes
- Loan repayments
- Software subscriptions
- Capital expenditure
Example prompt:
Create a 13-week cash-flow forecast using the Receivables, Payables, Payroll and Operating Expenses worksheets. Show the expected opening balance, cash inflows, cash outflows and closing balance for each week. Highlight any week where the projected balance falls below S$100,000.
Businesses can also use it for scenario analysis.
Example prompt:
Create three cash-flow scenarios: base case, optimistic case and conservative case. For the conservative case, assume customer payments are delayed by 30 days and monthly sales are 15% lower.
This helps business owners understand how changes in revenue, payment collection or expenses could affect their available cash.
4. Accounts Receivable and Overdue Invoice Analysis
An accounts receivable workbook may contain customer invoices, due dates, outstanding amounts and payment status.
GPT-5.6 can help categorise outstanding invoices into ageing groups such as:
- Current
- 1–30 days overdue
- 31–60 days overdue
- 61–90 days overdue
- More than 90 days overdue
Example prompt:
Analyse the Accounts Receivable worksheet and create an ageing report. Group outstanding invoices into current, 1–30 days, 31–60 days, 61–90 days and more than 90 days overdue. Show the total outstanding amount for each customer and identify the ten customers requiring the most urgent collection action.
It can also help prepare a collection list.
Example prompt:
Create a follow-up worksheet containing customers with invoices more than 30 days overdue. Include the customer name, invoice number, due date, amount, number of days overdue and recommended priority level.
This provides finance teams with a more organised way to manage customer collections.
5. Invoice and Payment Reconciliation
Reconciliation often involves comparing information from two separate sources.
For example:
- Sales invoices versus bank payments
- Supplier invoices versus accounting records
- License quantities versus vendor invoices
- Payment gateway transactions versus bank deposits
- Customer subscriptions versus billing records
GPT-5.6 can help compare datasets even when the references are not perfectly consistent.
Example prompt:
Compare the Invoice Register worksheet with the Bank Transactions worksheet. Match payments using invoice number, customer name, amount and payment date. Create separate lists for matched payments, unmatched invoices, unmatched bank transactions and possible partial payments.
For a cloud reseller or managed service provider, another useful example would be:
Compare the vendor subscription report with the customer billing report. Identify customers where the license quantity, product name, billing period or subscription status does not match.
This can reduce the amount of time employees spend manually checking transactions line by line.
6. Pricing and Profitability Analysis
Businesses can use GPT-5.6 to analyse whether products or services are being sold profitably.
A pricing workbook might include:
- Supplier cost
- Selling price
- Discount
- Gross profit
- Gross margin
- Sales volume
- Support cost
- Commission
- Renewal period
Example prompt:
Calculate the gross profit and gross margin for every product. Highlight products with a margin below 20%. Create a summary showing which products generate the most revenue and which generate the most profit.
GPT-5.6 could also support pricing simulations.
Example prompt:
Model the effect of increasing supplier costs by 8%. Show the new selling price required to maintain the current gross margin for every product.
Another scenario could be:
Create a sensitivity table showing the impact on monthly gross profit if prices are increased by 3%, 5% and 8%, assuming sales volume falls by 0%, 5% or 10%.
This helps management evaluate pricing decisions before implementing them.
7. Subscription and Recurring-Revenue Analysis
Subscription businesses need to monitor renewals, cancellations, expansion and recurring revenue.
GPT-5.6 can help analyse:
- Monthly recurring revenue
- Annual recurring revenue
- New subscriptions
- Renewals
- Upgrades
- Downgrades
- Cancellations
- Churn rate
- Revenue retention
- Upcoming renewal dates
Example prompt:
Analyse the Subscription worksheet and calculate monthly recurring revenue by product and customer. Identify subscriptions expiring within the next 30, 60 and 90 days. Highlight customers with a high renewal value that require early follow-up.
Another useful prompt would be:
Compare this month’s subscriptions with last month’s subscriptions. Identify new customers, cancelled subscriptions, quantity increases, quantity reductions and price changes.
For IT service providers, this could help reconcile Microsoft 365, Google Workspace, cybersecurity, backup and hosting subscriptions.
8. Inventory and Procurement Planning
Companies carrying physical stock can use GPT-5.6 to monitor inventory levels and purchasing requirements.
A workbook may contain:
- Product code
- Product description
- Current stock
- Average monthly sales
- Supplier lead time
- Minimum stock level
- Reorder quantity
- Unit cost
Example prompt:
Analyse the Inventory worksheet and identify products likely to run out within the next 60 days. Consider average monthly sales, current stock and supplier lead time. Recommend a reorder quantity for each affected product.
GPT-5.6 can also identify slow-moving stock.
Example prompt:
Identify products with no sales during the last 90 days or with more than six months of stock remaining. Calculate the total value of slow-moving inventory.
This can help businesses reduce stock shortages while avoiding unnecessary purchases.
9. Marketing Campaign Analysis
Marketing teams can combine advertising expenditure, leads and sales data in Excel.
GPT-5.6 can help calculate:
- Cost per lead
- Cost per acquisition
- Conversion rate
- Revenue per campaign
- Return on advertising spend
- Customer acquisition cost
- Revenue by marketing channel
Example prompt:
Compare the Google Ads, LinkedIn and Facebook campaign results. Calculate impressions, clicks, click-through rate, leads, conversion rate, cost per lead, customers acquired, revenue and return on advertising spend.
It can then explain the findings.
Example prompt:
Identify which campaigns should receive more budget, which should be improved and which should be stopped. Base the recommendation on customer acquisition cost, conversion rate and revenue generated.
This turns a spreadsheet from a reporting tool into a marketing decision-support system.
10. Customer Profitability Analysis
The customer generating the most revenue is not always the most profitable.
Some customers may require extensive support, receive large discounts or generate frequent service requests.
GPT-5.6 can help combine:
- Customer revenue
- Product cost
- Discounts
- Sales commission
- Support hours
- Service delivery cost
- Gross profit
- Gross margin
Example prompt:
Calculate the estimated profitability of every customer by subtracting product costs, discounts, commissions and support costs from revenue. Rank customers by total profit and profit margin.
Another prompt could be:
Identify high-revenue customers with below-average profitability. Explain the main reasons for the low margin, such as discounts, product mix or support requirements.
This can help account managers make better decisions about renewals, discounts and service levels.
11. Project Cost and Profitability Tracking
Professional services companies can use Excel to track project budgets, employee hours and expenses.
GPT-5.6 could help calculate:
- Budget consumed
- Hours used
- Remaining hours
- Labour cost
- External expenses
- Revenue recognised
- Estimated completion cost
- Project margin
Example prompt:
Compare planned hours with actual hours for every project. Identify projects that have used more than 80% of their allocated hours but are less than 70% complete.
Another example:
Estimate the final cost and profit margin for each active project based on the current spending rate and percentage of work completed.
This can provide early warning when projects are at risk of exceeding their budget.
12. Human Resources and Headcount Planning
HR teams can use GPT-5.6 to analyse non-sensitive workforce data such as:
- Department
- Job role
- Employment start date
- Salary band
- Leave balance
- Training completion
- Headcount
- Vacancy status
Example prompt:
Create a monthly headcount report by department. Show new hires, resignations, open positions and net headcount changes.
Another example:
Analyse annual leave balances and identify employees or departments with unusually high unused leave. Do not make assumptions about individual performance.
For workforce budgeting:
Forecast payroll costs for the next financial year based on current headcount, planned hires and an assumed 4% salary adjustment.
Companies should apply appropriate privacy and access controls when working with employee information.
13. Formula Creation and Troubleshooting
Many Excel users know the result they want but do not know which formula to use.
GPT-5.6 can translate a plain-language request into an appropriate formula.
Examples include:
Create an XLOOKUP formula that returns the customer’s payment terms using the customer ID in cell A2.
Write a SUMIFS formula that totals revenue for Singapore customers during July 2026.
Create a formula that returns “Renewal Due” when the expiry date is within the next 60 days.
Use FILTER to display only invoices that are overdue and have an outstanding balance above S$5,000.
Create a formula that calculates the number of working days between the order date and delivery date, excluding weekends and Singapore public holidays.
Use LET to simplify this long formula and make it easier for another employee to maintain.
GPT-5.6 can also explain existing formulas.
Example prompt:
Explain the formula in cell G24 in plain English. Identify every cell and worksheet it depends on, and explain what could cause it to return an incorrect result.
For an error:
The formula returns #N/A for some customers. Investigate whether the problem is caused by missing records, extra spaces, number formatting or an incorrect lookup range, and recommend a fix.
14. Data Cleaning and Standardisation
Business data is often messy, especially when it has been copied from multiple systems.
Common problems include:
- Duplicate records
- Inconsistent date formats
- Customer names written differently
- Numbers stored as text
- Empty rows
- Leading or trailing spaces
- Mixed currencies
- Inconsistent capitalisation
- Missing values
Example prompt:
Review the Customer Data worksheet for quality issues. Identify duplicate customers, blank mandatory fields, invalid email addresses, inconsistent country names and customer IDs stored in different formats.
Another useful prompt:
Standardise all company names by removing unnecessary spaces and inconsistent punctuation. Do not merge records automatically. Create a separate worksheet showing possible duplicates for review.
This helps ensure that the original data is not changed without oversight.
15. Dashboard Creation
GPT-5.6 can help transform detailed worksheets into a management dashboard.
A dashboard could include:
- Monthly revenue
- Gross profit
- Cash balance
- Outstanding invoices
- New customers
- Customer churn
- Sales pipeline
- Service performance
- Project status
- Inventory alerts
Example prompt:
Create an executive dashboard using the existing workbook. Include monthly revenue, gross margin, overdue receivables, new customers and recurring revenue. Add charts showing the 12-month trend and comparisons against budget.
GPT-5.6 can also recommend the most suitable chart for each metric.
For example:
- Line charts for monthly trends
- Bar charts for comparisons
- Waterfall charts for profit movements
- Stacked charts for product mix
- Heat maps for performance by region
- KPI cards for headline figures
Because GPT-5.6 has improved spreadsheet presentation and layout capabilities, it can help produce worksheets that are not only functional but also easier for managers to read and use.
16. Spreadsheet Auditing and Error Detection
Complex workbooks can contain hidden errors that affect important business decisions.
GPT-5.6 can help review a workbook for:
- Broken cell references
- Inconsistent formulas
- Hard-coded numbers inside formulas
- Circular references
- Missing values
- Unusual changes in assumptions
- Totals that do not reconcile
- Formulas overwritten by fixed values
- Inconsistent date ranges
Example prompt:
Audit this financial workbook. Identify inconsistent formulas, hard-coded assumptions, broken references, formulas that differ from surrounding cells and totals that do not reconcile. Create an Issues worksheet listing the cell, problem, potential impact and recommended action.
Another example:
Compare the current workbook with last month’s version. Identify changes to formulas, assumptions, source data and worksheet structure.
This is especially valuable when several employees contribute to the same workbook.
17. Scenario and Sensitivity Analysis
GPT-5.6 can help businesses evaluate several possible outcomes before making a decision.
For example, a company might want to understand what happens when:
- Sales increase or decrease
- Supplier costs rise
- Currency exchange rates change
- Customer payment periods become longer
- Headcount increases
- Rental costs rise
- Prices are adjusted
- Customer churn changes
Example prompt:
Build a scenario analysis for the next 12 months. The base case should use the current forecast. The optimistic case should assume 15% revenue growth and stable costs. The conservative case should assume revenue falls by 10%, supplier costs rise by 8% and customer payments are delayed by 15 days.
Another example:
Create a sensitivity table showing how gross profit changes when sales growth ranges from negative 10% to positive 20% and gross margin ranges from 15% to 30%.
This enables management to understand both the potential return and the downside risk of a decision.
18. Turning Excel Analysis into Management Commentary
GPT-5.6 can help convert spreadsheet results into a written business explanation.
Example prompt:
Review the completed monthly report and write a 300-word executive summary. Explain the key revenue, profitability, cash-flow and customer trends. Focus on material changes and recommend three management actions.
It can also tailor the explanation for different audiences:
Rewrite the analysis for a finance manager and include detailed variances.
Rewrite the analysis for the managing director using plain English and focus on the five most important findings.
Turn the findings into a board-level summary with key risks, decisions required and recommended next steps.
This removes the need to manually copy numbers from Excel into a separate report.
Creating Repeatable Excel Workflows
One of the more useful developments is the ability to create reusable spreadsheet workflows.
ChatGPT for Excel and Google Sheets supports Skills, which act as reusable playbooks containing instructions, formatting standards and expected output structures. OpenAI also provides default Skills for financial modelling and corporate finance formatting.
For example, a company could create a reusable monthly reporting process that instructs ChatGPT to:
- Import the latest sales data.
- Preserve the existing worksheet structure.
- Refresh formulas and charts.
- Compare results against the previous month.
- Highlight material variances.
- Update the management dashboard.
- Produce an executive summary.
- Check the workbook for errors before completion.
This can improve consistency when the same report must be prepared every week or month.
ChatGPT Work can also create or edit spreadsheet files using instructions, source material or an existing company template. Users can specify which formulas, layouts, branding elements or table structures must remain unchanged.
Example of an End-to-End Excel Request
A complete business prompt could look like this:
Review the attached monthly management workbook. Preserve the existing worksheets, formulas and formatting. Import the latest sales and expense data, update the dashboard and compare July 2026 against June 2026 and July 2025. Identify material changes in revenue, gross margin, operating expenses, outstanding invoices and cash flow. Check for broken formulas or inconsistent calculations. Add a Management Summary worksheet containing the ten most important findings, key risks and recommended actions. Clearly identify any assumptions or data-quality problems that require human review.
This type of instruction combines data preparation, formula checking, analysis, presentation and reporting into one workflow.
Important: Human Review Is Still Necessary
GPT-5.6 can make Excel work faster, but businesses should not treat every generated formula or conclusion as automatically correct.
Users should review:
- Source data
- Financial assumptions
- Formulas
- Cell references
- Tax calculations
- Forecasting methods
- Accounting treatment
- Final charts and summaries
OpenAI also recommends reviewing formulas, source data and workbook changes before saving, sharing or relying on the completed spreadsheet.
The best approach is to use GPT-5.6 as an Excel copilot that handles repetitive work, proposes analyses and highlights possible issues, while employees remain responsible for approving the final output.
Why GPT-5.6 Could Change How Businesses Use Excel
Traditionally, advanced Excel work required employees to understand formulas, PivotTables, data models and financial modelling techniques.
GPT-5.6 lowers that barrier by allowing employees to describe what they need in normal language. A sales manager can request a revenue dashboard, a finance executive can request a cash-flow forecast, and an operations manager can request a service-performance report without building every calculation manually.
The result is not the replacement of Excel expertise. Instead, GPT-5.6 can help experienced users complete sophisticated work faster while enabling less technical employees to analyse data more confidently.
For small and medium-sized businesses, this may be one of GPT-5.6’s most practical benefits: turning existing Excel files into more useful tools for reporting, forecasting and decision-making.
