AI 現金流代理：避免在生意做不下去前先耗盡現金流

Cash flow failure is counterintuitive because profit is an accounting concept that recognizes revenue when it is earned, not when it is received. A business that invoices enterprise customers on 60-day terms, pays suppliers on 30-day terms, and carries three months of inventory has a structural cash timing mismatch regardless of margins. Example: a services business wins a £200,000 contract. Revenue hits the P&L immediately, the team is hired, the work is delivered, costs are paid, and the invoice is raised. Sixty days later the cash arrives — but by then, two months of salaries, supplier payments, and overheads have been paid from existing cash.

A 13-week rolling cash flow forecast is the standard tool for active cash management because it balances two competing needs: enough granularity to make decisions (weekly rather than monthly) and enough horizon to act (13 weeks is sufficient time to address most working capital problems before they become crises). Monthly cash projections in management accounts are too coarse and too late. The forecast shows weekly cash inflows by revenue stream, weekly outflows by category, closing balance each week, and a minimum cash threshold — making danger weeks visible weeks or months before they arrive.

The four main levers are: collect receivables faster (reduce average debtor days, chase outstanding invoices systematically, introduce early payment discounts), pay payables slower (negotiate extended payment terms with suppliers, use full payment terms available rather than paying early), reduce inventory or work-in-progress (identify slow-moving stock, reduce buffer inventory, bill clients sooner), and increase prepayments (shift customers from invoice-after-delivery to deposit-plus-final-payment structures). A business with £500K of outstanding receivables at 60-day average terms can release £83,000 of cash by moving to 45-day terms — without a single new customer.

Cash management requires three scenarios: the base case models the current trajectory with current revenue, costs, and payment terms. The downside case models revenue at 20-30% below base case, reflecting a realistic bad quarter rather than catastrophic one. The improvement case models working capital actions implemented — improved debtor days, extended payable terms, activated levers. Each shows cash runway in weeks and the specific month where intervention would be required. Knowing you have 22 weeks of runway in a downside scenario gives 22 weeks to act — discovering you have four weeks when the bank balance falls is 18 weeks too late.

A cash flow forecast should be reviewed weekly, every Monday morning. The weekly monitoring framework includes the five numbers to check, variance analysis when actuals differ from forecast, escalation triggers (the cash balance at which specific actions must be taken), and decision trees for responding to different scenarios. A forecast is only useful if it is maintained and reviewed regularly. The forecast is the map — the monitoring framework is the discipline of consulting it regularly rather than once. For cash-stressed businesses, this review should start immediately.