For lenders
Every R01 return is a payment you already counted as collected. Per-call balance checks add cost on every retry, and still clear debits that come back.
New from Pave · ACH Risk Score
Cut ACH returns before they happen.
Live balance checks don't predict returns. Pave ACH Risk Score reads the cashflow signals that actually move with risk and scores every ACH before it's sent. Real-time, explainable, and tunable by attempt type.
One of the Pave Cashflow Scores, the family of risk-event-tuned scores built on Pave's enriched cashflow data.
Running in production at lending, EWA, and fintech programs
.avif)
.avif){kind=logo}
.avif)
.avif)
%20(4).avif){kind=logo}
.svg){kind=logo}
Scoring ACH attempts across consumer cash advance, earned wage access, and small-dollar lending programs.
Where Returns Come From
A balance check tells you the money is there now, not whether it'll settle in a few days.
Most ACH risk controls answer the wrong question. A balance check confirms funds at the moment of submission. But a user can appear to have the funds, say $300 at 2pm, and still return because their balance is back to $0 by the time the debit settles. By then the customer has been notified and the cost is on your books. The decision had to happen before you sent the entry.
For EWA & Cash advance
Return rates drift toward the NACHA threshold as your new-user mix grows. A static rule deployed once drifts out of compliance.
For fintechs & BaaS
Account validation catches ownership, not behavior. Sponsor banks ask for return-rate controls every quarter.
Pave ACH Risk Score
The probability the next debit returns, scored before you send it.
Pave ACH Risk Score reads balance trajectory, paycheck cadence, and recent return history, then returns the odds the attempt comes back. It learns from real outcomes across our network. Every debit our lenders run, settled or returned, trains the score, which updates as each borrower's balance and deposits change.
A probability, not a snapshot.
Balance checks tell you funds are present right now. Pave returns the probability the debit comes back, read from cashflow behavior the snapshot can't see.
Tunable by attempt type.
Set separate operating points for due-date submissions and overdue retries (different risk, different cutoffs), and re-tune as your population shifts.
A quick integration, not a data project.
The score runs on enriched cashflow data from the bank account, the same data your team already uses to underwrite. One API call, no new pipeline to stand up.
How It Works
From attempt to decision in one call.
1
Send the attempt
Score it in-line before submission, or batch a file before it goes out.
2
Pave scores it
Against the bank cashflow data: a probability the attempt returns, with the factors behind it.
3
Act on the decision
Submit, hold, or retry against the thresholds you set.
Proof
An earned wage access provider got back under the NACHA threshold, and stopped paying for balance checks.
Their ACH return rate had climbed past 35%, more than 2x NACHA's 15% line, and a balance check on every retry was both expensive and inconclusive. They deployed Pave ACH Risk Score with separate thresholds for due-date and overdue attempts.
35% → 13%
ACH return rate, back under NACHA's 15% threshold
81.2% → 91.9%
Collection success: reducing returns didn't cost recovery
49.3%
Of all returns concentrated in the top-risk decile, 4.3x baseline
The score replaced the live balance check as the decision logic on the retry path, eliminating the balance check, its per-retry cost, and the overdraft fees that came from retrying against stale funds. It validated in the mid-80% range on live traffic, and the integration was a configuration change, not a new pipeline.
Read the full case study→Where It Fits
One score in a family, built on one cashflow data layer.
Pave is the AI credit infrastructure for lenders. ACH Risk Score is one of the Pave Cashflow Scores: risk-event-tuned scores that all read from the same enriched cashflow data, so the signal you score an ACH retry with is the same signal behind every other Pave score.
LAYER 1
Transaction Enrichment
Cleans, standardizes, and categorizes raw bank transaction data into consistent, reliable signals.
LAYER 2
Attributes
Thousands of derived cashflow features (income, spending, and repayment patterns), ready to plug into any model.
LAYER 3
Attributes
Thousands of derived cashflow features (income, spending, and repayment patterns), ready to plug into any model.
YOU ARE HERE - ACH RISK SCORE
UPCOMING
Credit Risk Agent
Agentic credit modeling on top, so your team can build and own custom models for any risk event, in days. You keep the model and the judgement; the agent does the mechanical work.
Use Cases
One score for every ACH decision.
Keep your return rate under the NACHA threshold as your book grows.
.svg)
Separate operating points for due-date submissions and overdue retries
Concentrate returns in the top-risk decile so operations has a precise lever
Re-tune thresholds as your new-user mix shifts, holding compliance margin through portfolio change
Recover compliance without throttling advance volume
Release the credit limit the moment they pay, not days later.
Score ACH return risk at the moment of payment, so you can release the limit immediately for payments you can confidently clear
Cut the held-funds window that drives support tickets, confusion, and churn
A servicing and compliance win that doesn’t touch your underwriting policy
Built for card programs, neobanks, and pay-down-to-free-up-limit flows
Cut loan-payment returns and stop paying per-call for balance checks.
Score each debit and retry for return risk before the entry is sent
Replace a per-call balance-check vendor with a score that reads the cashflow data you already have
Explainable factors on every score for ops, audit, and sponsor-bank review
Built for personal loans, BNPL, POS finance, and small-dollar lending
Pave Index · Vol. 01
2026 Industry Report
The State of ACH Risk
Return codes, predictability, and what well-run programs actually hit. Across $400B in flows.
Coming this summer
Pave 2026 · Vol. 01
GET A DEMO
Get Your Personalized Demo
Our team will walk you through how lenders and fintechs use Pave to analyze cashflow data, score applicants with confidence, and grow sustainably—without increasing risk.