DEMO / 04 · Demo ready
Cashflow invoice lab
Synthetic receivables lab: three-way matching, duplicate invoice checks, payment-delay baselines and a 13-week cash forecast with intervals.
What decision is supported
Which invoices to release for payment, which to hold for a mismatch or a suspected duplicate, and how much cash the business should expect to collect in each of the next thirteen weeks.
Who uses it
An accounts-payable or receivables analyst at a fictional distributor, and a finance lead who reads the forecast interval rather than a single number.
What data enters
Authored synthetic purchase orders, goods receipts, invoices and payment records for fictional counterparties. No real invoice, vendor or bank record is used.
What is computed
Three-way match results between order, receipt and invoice with tolerance rules, duplicate detection on normalised invoice keys, payment-delay baselines per counterparty from history, and a 13-week collections forecast with a stated interval from an empirical residual distribution.
What action is suggested
Release, hold with reason, or flag as a probable duplicate; a weekly cash range rather than a point estimate.
What evidence supports it
Forecast accuracy is reported only on the synthetic fixture inside the repository, labelled demo. It is not a claim about any real ledger.
No metric is claimed for this project. Anything computed by the repository on its synthetic fixture carries the label demo.
What fails or is uncertain
- Payment behaviour in the fixture is generated from stated distributions; real counterparties are less regular.
- The forecast is a baseline method with intervals, chosen for inspectability, not the best possible forecaster.
- Currency is a demo unit; no exchange rates or tax rules are modelled.