What you need
- Opening cash balance
- AR aging and AP aging (replace customer and vendor names with codes)
- Payroll calendar and amounts
- Debt service, rent, taxes and other recurring payments
- Known one-time receipts and payments
Steps
- Collect the inputs and replace names with codes like "Customer A".
- Paste the prompt. Answer the AI's clarifying questions before it builds.
- Build the model in Excel from the structure it gives you, keeping assumptions in separate cells.
- Each week, roll the forecast forward and compare actual cash to the forecast.
The prompt
Help me build a 13-week cash flow forecast for a [industry] company. Opening cash as of [date]: $[amount] Minimum cash we want to keep: $[amount] Line of credit: $[limit] with $[drawn] drawn Data: [paste AR aging by bucket, AP aging by due date, payroll calendar, recurring payments, one-time items] Before building, ask me up to five clarifying questions about anything missing or unclear. Then lay out a week-by-week model with: - Receipts: collections from each AR aging bucket using explicit collection-rate assumptions, plus other receipts - Disbursements: payroll, AP by due week, rent, debt service, taxes, other - Net cash flow, beginning and ending cash, and line of credit availability - A flag in any week where ending cash falls below the minimum List every assumption separately so I can change it, and give the Excel formulas for the structure.
Review before you use it
- Week 1 opening cash ties to the bank balance.
- Collection rates are based on your actual collection history, not optimism.
- Payroll and tax dates match the real calendar, including bonus or quarterly payments.
- A weekly forecast-versus-actual comparison is in place.
Never paste into an AI tool
- Customer and vendor names (use codes)
- Bank account numbers
- Use a business or enterprise AI plan that does not train on your data, and follow your company's AI policy.
Educational content, not professional advice. AI output can be wrong; review it before relying on it. Disclaimer