A 30/60/90 receivables forecast that updates itself

Abstract

Cash-flow visibility usually depends on a manual spreadsheet that's already stale the day it's built. We built an agent that cross-references each customer's payment history with their open invoices to produce a rolling 30/60/90 collections forecast, with a probability of payment per customer.

The problem

Treasury decisions ride on knowing what's likely to land and when, but the forecast is a hand-built spreadsheet maintained by one person and out of date almost immediately. The result is either over-caution or nasty surprises.

What we built

• Ingestion of payment history and open invoices from Holded.
• A model of each customer's payment behaviour over time.
• A rolling 30/60/90 collections forecast with a per-customer probability.
• Automatic refresh as invoices and payments change.

How it works

• The agent reads historical payments and current receivables.
• It estimates when each open invoice is likely to be collected.
• It rolls those into a 30/60/90 view with confidence per customer.
• The forecast updates itself as the underlying data moves.

Result

Treasury gets a live, defensible collections forecast instead of a stale spreadsheet — earlier warning on slow payers and a clearer picture of the weeks ahead.