Know which invoices will pay late. Chase them before they do.
CashPilot predicts late payment for every open invoice from your customers' history, ranks your collectors' day by money at risk, and runs evidence-based reminder sequences automatically, so the team works the invoices most likely to slip without adding headcount.
Priority worklist live ranking
- Orchard Lane NurseriesINV-10484 · 34,543critical96%33,161
- Goldleaf AccountingINV-10250 · 26,231critical95%24,919
- Crestline SolarINV-10691 · 23,799high71%16,897
- Dunmore TextilesINV-11318 · 40,932medium38%15,554
- Granite LogisticsINV-10041 · 22,233low12%2,668
Risk value = balance × P(late). The biggest balance is not the first call.
Imports from any accounting system · connects to the payment rails you already use
One system from prediction to reconciled cash.
Built for B2B companies with 50 to 5,000 open invoices who have outgrown spreadsheets and aging reports. Every piece below is in the live demo.
Payment prediction, per invoice
A gradient-boosted model trained on your own ledger scores every open invoice with the probability it pays late and its most likely pay date. 24 features, retrained in one click, calibration and feature importance you can inspect.
Feature importance, demo ledger · accuracy 79% vs 52% baseline · AUC 0.83
Risk-ranked worklist
Collectors open one list ordered by balance × P(late). Bulk remind, assign, pause or create call tasks.
Reminder sequences
Pre-due courtesy, then escalating touches. Paused on promise or dispute.
13-week cash forecast
Predicted pay dates roll up into weekly expected and promised cash, with a conservative haircut past 60 days. The number your CFO actually asks for.
Pay links and bank feeds
Customers pay by card through Stripe Checkout from the reminder. Statements arrive as camt.053, MT940, OFX or CSV, or via Plaid and GoCardless, and match themselves to invoices.
Customer portal
Every reminder links to a live statement where customers confirm, query or pay.
Receivables KPIs
DSO, best-possible DSO, ADD and CEI to Credit Research Foundation definitions.
Demo workspace figures, to Credit Research Foundation definitions.
API and webhooks
Scoped keys, rate limits with standard headers, signed events. Your ledger stays the source of truth.
Governance built in
Four roles, an audit log with before and after values on every change, per-organisation send windows and time zones, and a release checklist you can read inside the product.
Every route declares its minimum role. Cross-tenant access is a 404.
From CSV to reconciled cash, in five steps.
Live in an afternoon. Each step below is a real screen from the demo workspace.
Import and score
Upload customers and invoices from Xero, QuickBooks, Sage, NetSuite or any CSV. Paid history trains the model on day one; every open invoice gets a probability and a predicted pay date.
Work the ranked list
Collectors see the invoices most likely to slip, filtered by collector, risk band and days overdue, with promises and disputes in view. Bulk actions in one click.
Let the reminders run
Sequences send inside your window with the invoice, statement and pay link attached, escalating in tone. They pause on a promise or dispute and resume when it breaks or resolves.
Forecast the cash
Predicted pay dates roll up into a weekly forecast, split into expected and promised, exported to CSV or read through the API.
Reconcile when money lands
Card payments post on signed Stripe webhooks. Bank lines match by reference, amount and payer; high-confidence matches apply themselves, the rest queue with suggestions.
Every day of DSO is cash sitting in someone else’s bank.
Move the sliders. The estimate uses the standard working-capital identity: one day of DSO equals one day of credit sales.
Estimate only. Cash released = revenue ÷ 365 × days of DSO removed. Hours assume six hours a week of manual chasing per $1M of credit sales. Your results depend on your ledger; the trial shows them on your own data.
See it on your ledgerGrounded in published research, not folklore.
The model, the ranking rule and the reminder cadence each trace back to a study you can read. See the evidence →
Aggregated customer payment history lifts invoice-level prediction accuracy far above invoice fields alone. Ranking collectors by value × probability saved a bank about $1.75M a month in simulation.Zeng, Melville, Lang, Boier-Martin, Murphy · KDD 2008 · Appel et al. 2020
An invoice chased in its first week overdue is about two and a half times more likely to be paid than one first chased after 30 days. CashPilot's default sequence contacts at +1 and +7.Atradius Payment Practices Barometer
Field experiments find plain, personalised reminders outperform elaborate ones, while social-norm appeals do not move payment at all. Our templates are short, specific and carry the evidence.Behavioural field experiments on payment reminders
Built like it has to hold money.
Every release passes an automated acceptance checklist before it ships. The current one runs 26,863 checks, all green, and you can read it inside the product. Security details →
Priced on open invoices, not on seats.
Free for up to 25 open invoices. Paid plans from $79 a month include automated sequences, the portal, pay links and bank import. No per-reminder or per-seat fees.
Starter
For finance teams getting off spreadsheetsUp to 250 open invoices · 3 users
- Everything in Free
- Automated reminder sequences
- Customer portal and statements
- Stripe pay links
- Bank statement import (camt.053, MT940, OFX, CSV)
- Email support
Growth
For teams collecting at scaleUp to 1,500 open invoices · 10 users
- Everything in Starter
- Plaid and GoCardless bank feeds
- REST API and signed webhooks
- Daily collector digests
- Custom sequences per segment
- Priority support
Scale
For groups and shared service centresUp to 6,000 open invoices · 25 users
- Everything in Growth
- SSO and audit exports
- Sandbox workspace
- Named success manager
- 99.9% uptime SLA
Free · $0
Up to 25 open invoices · 1 user. For founders chasing their own invoices. Payment prediction and ranked worklist, 13-week cash forecast, Reminders sent one at a time, not as a sequence, CSV import and export.
14-day trial on every paid plan · Prices in USD, exclude tax · Full comparison, plan recommender and FAQ →
Frequently asked.
How long does it take to get started?
Does it replace our accounting system?
How accurate is the late-payment prediction?
Will customers receive reminders we did not approve?
How do card payments and bank feeds work?
Where is our data stored, and who can see it?
Can we run it ourselves?
What does the trial include?
Stop financing your customers’ cash flow.
Import a CSV and see your own ranked worklist and 13-week forecast in about ten minutes. Or click around the live demo workspace with 18 months of ledger history.