Which week will the cash run short, and can you see it a month and a half ahead?

The position three months out is a correct number, and almost useless. Money doesn't run out at the end of the quarter. It runs out in one specific week, between payroll and a client who pays at ninety days.

The forecast takes the same three terms as the Bank screen, the balance, receipts and payments, and spreads them over thirteen weeks.

The bars don't stand on zero but on your floor: the balance below which the company considers itself in trouble. In the demo, a hundred and eighty thousand dirhams. By default the floor is zero: the program doesn't invent your cushion.

The low point is week six: two hundred and sixteen thousand, just thirty-six thousand above the floor. I click the week, and it opens down to the documents: payroll, social security, a client invoice. An argument nobody can check convinces nobody.

The screen prints the limits of its own accuracy: money on the account that no document explains, and debts that fall due after the horizon.

Take the scenario "the main debtor pays a month late": the low point drops below the floor, and a shortfall week and a funding need appear. The scenario changes no data. It's a question, not a claim.

The Freeze button stores a version of the forecast. And the Accuracy tab compares past versions with what was really on the account, and says plainly whether the forecast was optimistic. This is a page you can show your bank.

The week cash runs short, a month and a half ahead, not the day before.