▤ Insights · payment operations

Payment capture & allocation.

How customer payments are captured, remittance advice is gathered, and incoming funds are matched and allocated — from simple bank transfers to AI-driven multi-currency reconciliation.

What this covers

The end-to-end landscape of how customer payments are captured, remittance advice is gathered, and incoming funds are matched and allocated — from simple bank-transfer acknowledgements to AI-driven multi-currency, high-volume reconciliation engines.

Who it's for

O2C directors, GBS/SSC architects, treasury leads, SAP S/4HANA implementers, and AR transformation teams building or rationalising their payment stack.

How to read

Cards are grouped by function, with a volume indicator for suitability. Pros and cons reflect real-world implementation experience, not vendor claims.

01

Remittance advice capture

How customers tell you what they paid — and why
Low volume / SME
Mid volume
High volume
Complex / multi-entity
Remittance · manual

Email / PDF remittance

Customer emails a PDF, Excel, or scanned document listing invoice numbers and amounts paid. The most common format for SME and mid-market B2B globally. Requires human interpretation or OCR extraction.

LowMid
Strengths
  • Universal — every customer can do it
  • No integration needed
  • Captures complex deductions/disputes inline
Limitations
  • Labour-intensive at scale
  • No standard format → data-quality issues
  • Latency 24–72h before allocation
OCRIDPAI extractionemail parsing
Remittance · portal

Customer self-service portal

Customers log in to a supplier portal (SAP BTP, Billtrust, HighRadius, Invoiced) and explicitly match their payment to open invoices before or at point of payment. Near-perfect allocation data.

MidHigh
Strengths
  • Structured, clean data at source
  • Enables self-service dispute filing
  • Reduces AR team workload dramatically
Limitations
  • Portal adoption requires customer effort
  • Build & maintenance cost
  • Large customers resist non-integrated flows
SAP BTPHighRadiusBilltrustInvoiced
Remittance · EDI / B2B

EDI 820 / EDIFACT REMADV

Structured electronic remittance transmitted via EDI or AS2. Standard in retail, FMCG, automotive, and pharma. Carries invoice-level payment detail, deduction codes, and chargeback references in machine-readable format.

HighComplex
Strengths
  • Fully automated end-to-end
  • Industry standard — no mapping ambiguity
  • Real-time or near-real-time ingestion
Limitations
  • Onboarding effort per trading partner
  • Requires EDI middleware (Seeburger, MFT)
  • Legacy customers may not support it
EDI 820EDIFACTAS2SeeburgerOpenText
Remittance · AI/IDP

Intelligent document processing

AI/ML models (Google Document AI, AWS Textract, Rossum, ABBYY, HighRadius) ingest unstructured PDFs, Excel files, and images, and extract invoice references, amounts, and deduction codes automatically — replacing manual keying.

MidHighComplex
Strengths
  • Handles diverse formats without pre-mapping
  • Confidence scoring enables STP routing
  • Continuous learning improves over time
Limitations
  • Training data & validation required upfront
  • Low-confidence docs still need human review
  • Cost-per-document model
RossumABBYYTextractDocument AIHighRadius
Remittance · API

Real-time API remittance

Large enterprise customers transmit remittance data via REST or SOAP APIs at point of payment — common with Tier 1 retailers and automotive OEMs using supplier networks (Coupa Pay, Ariba, Tungsten/OB10). Instant allocation.

HighComplex
Strengths
  • Zero latency — allocate at time of payment
  • Eliminates reconciliation backlog
  • Supports complex deduction structures
Limitations
  • Requires customer IT participation
  • Platform-specific (Coupa, SAP Ariba)
  • Only feasible for top-tier customers
Coupa PayAriba NetworkTungstenREST API
Remittance · banking

Structured payment references (IBAN + ref)

Assign unique virtual IBANs or structured creditor references (ISO 11649, Betalingskenmerk NL, BankgiroNr SE) per invoice or customer. Bank feeds deliver pre-matched data with no separate remittance required. Dominant in NL, BE, Nordics.

LowMidHigh
Strengths
  • No separate remittance process needed
  • High STP rate — often 95%+
  • Works for any payment volume
Limitations
  • Requires customer discipline on reference
  • Multi-invoice payments break the model
  • Virtual IBAN issuance cost
Virtual IBANISO 11649CAMT.054MT940
02

Payment capture channels

How funds actually arrive — instruments & rails
Capture · banking

SWIFT / international wire transfer

MT103 (single credit transfer) and ISO 20022 pain.001 cross-border payments. Standard for large B2B cross-border. Bank statement reconciliation via MT940/942 or CAMT.053/054 feeds.

MidHighComplex
Strengths
  • Global reach, any currency
  • ISO 20022 carries rich remittance data
  • Audit trail and irrevocability
Limitations
  • 1–3 day settlement
  • FX conversion costs
  • Limited remittance in MT103 (140 chars)
SWIFTMT103ISO 20022CAMT.053
Capture · SEPA

SEPA credit transfer & direct debit

SCT (1 day) and SCT Inst (10 seconds) for push payments. SDD Core/B2B for pull. The backbone of EUR B2B payments across 36 SEPA countries. Structured creditor reference enables high STP rates.

LowMidHigh
Strengths
  • Instant SCT Inst — real-time cash
  • SDD eliminates collection effort
  • Low cost, high reliability
Limitations
  • EUR-only (no multi-currency)
  • SDD mandate-management overhead
  • B2B SDD still limited bank support
SEPA SCTSCT InstSDD Corepain.008
Capture · PSP

Payment service providers (PSP)

Stripe, Adyen, Mollie, Worldline, Braintree aggregate card, wallet, BNPL, and local payment methods. Provide unified reconciliation reports and webhooks. Critical for B2C and digital-first B2B.

LowMidHigh
Strengths
  • Single API for 100+ payment methods
  • Webhook-driven real-time capture events
  • Built-in fraud, chargeback handling
Limitations
  • Settlement T+1 to T+3 (varies)
  • PSP reconciliation vs bank-statement gap
  • Per-transaction fee model at scale
StripeAdyenMollieWorldlinewebhooks
Capture · open banking

Open banking / account-to-account

PSD2-enabled A2A payments (Tink, Token.io, TrueLayer, iDEAL 2.0, Bancontact) let customers pay directly from their bank with real-time confirmation. Eliminates card rails. Fast-growing in NL, UK, DE.

LowMidHigh
Strengths
  • Real-time final — no chargebacks
  • Lower cost than card (no interchange)
  • Rich payment-initiation data
Limitations
  • Geographic coverage still uneven
  • Refund flows more complex than cards
  • Consumer-awareness gaps
PSD2iDEAL 2.0TinkTrueLayerPISP
Capture · collections

ACH / BACS / direct debit

Batch pull payment rails: ACH (US), BACS/Faster Payments (UK), Autogiro (SE), CORE (FR). Used for recurring B2B and subscription billing. Mandate-based with a 2–3 day float in older schemes.

MidHigh
Strengths
  • Predictable cash flow for subscriptions
  • Reduces DSO on recurring invoices
  • Batch-processing efficiency
Limitations
  • Return/reversal window (up to 8 weeks SDD)
  • Mandate-management burden
  • Not suitable for one-off or variable amounts
ACHBACSFaster PayAutogiro
Capture · cheque / virtual

Cheque & virtual card (vCard)

Physical cheques are declining globally but still active in US/CA. Virtual cards (single-use Mastercard/Visa issued by the buyer's treasury) are increasingly used by US corporates — captured via standard card rails but with embedded remittance in card data.

LowMid
Strengths
  • vCard: remittance encoded in Level 3 data
  • Familiar to legacy US/CA buyer base
  • vCard earns interchange rebate for buyer
Limitations
  • Cheques: manual, slow, fraud risk
  • vCard: high merchant fee (1.5–3%)
  • Acceptance not universal
Virtual CardLevel 3 datavCard STP
03

Payment matching & allocation engines

Connecting money received to the AR ledger
The core O2C challenge: payment matching is where most AR transformation value is unlocked. A best-in-class operation targets >95% STP (straight-through processing) — the system allocates automatically without human touch. Below 70% STP is a burning platform.
Matching · rule-based

Deterministic rule engine

Sequential matching rules: exact amount + invoice ref → partial amount matching → customer + amount → amount-only tolerance bands. Native in SAP (F-28, F-32), Oracle AR, and most ERP systems. Fast and predictable for clean data.

LowMid
Strengths
  • Transparent — explainable decisions
  • No model training needed
  • SAP standard config, low cost
Limitations
  • Brittle with dirty/incomplete references
  • STP rate typically 50–75% max
  • Rule maintenance at scale is complex
SAP F-28Oracle ARMS Dynamics
Matching · AI/ML

AI-powered cash application

HighRadius, Billtrust, YayPay, Esker, and SAP FSCM AI use ML models trained on historical match patterns to handle fuzzy references, partial payments, combined invoices, and multi-currency. STP rates of 85–97% achievable.

HighComplex
Strengths
  • Handles ambiguous, partial, combined payments
  • Self-improves with each correction
  • Processes remittance PDFs natively
Limitations
  • Requires data history for model training
  • Licence cost — ROI calc needed
  • Black-box risk for audit teams
HighRadiusBilltrustEskerSAP FSCMYayPay
Matching · bank integration

Bank statement auto-import (CAMT/MT940)

Automated import of bank statements (CAMT.053 daily, CAMT.054 intraday, MT940/942) into ERP. SAP Electronic Bank Statement (EBS) or equivalent maps transactions to GL and triggers AR clearing rules without manual intervention.

MidHigh
Strengths
  • Near real-time with intraday CAMT.054
  • Single source of truth — bank-confirmed
  • Standard SAP config, no custom dev
Limitations
  • Reference quality depends on customer
  • Multi-bank aggregation adds complexity
  • Intraday CAMT requires an extra bank fee
CAMT.053CAMT.054SAP EBSMT940TMS
Matching · treasury

Treasury management system (TMS) integration

Enterprise TMS (Kyriba, ION, SAP TRM, Reval) connects bank feeds, FX positions, and AR/AP clearing in a single platform. Critical for multi-currency, multi-entity, multi-bank treasury operations with netting and pooling.

HighComplex
Strengths
  • Cash visibility across all entities/banks
  • FX-hedge alignment with AR exposure
  • Intercompany netting reduces settlement volume
Limitations
  • High implementation cost
  • AR–TMS data-sync latency
  • Over-engineered for simpler organisations
KyribaIONSAP TRMReval
Matching · exceptions

Deduction & short-pay management

Customers (especially retail & FMCG) routinely deduct trade promotions, logistics penalties, damage claims, or pricing disputes. Specialist tools (HighRadius Deductions, Model N, Esker) capture, code, and route these for resolution with validity scoring.

HighComplex
Strengths
  • Structured dispute-lifecycle management
  • Auto-validation against promo agreements
  • Recovers more invalid deductions
Limitations
  • Requires trade-promo data integration
  • Cross-function (Sales, Finance, Ops)
  • High implementation complexity
DeductionsHighRadiusModel NEskerTPM
Matching · RPA

RPA-assisted cash application

UiPath, Automation Anywhere, or SAP Build bots navigate bank portals, download statements, extract email remittances, and key data into ERP — bridging the gap where API/EDI integration is not possible.

MidHigh
Strengths
  • No system changes required at source
  • Quick to deploy for specific pain points
  • 24/7 operation, no staffing dependency
Limitations
  • Brittle — breaks on UI changes
  • Not a strategic solution
  • Maintenance cost underestimated
UiPathAASAP BuildPower Automate
04

End-to-end process flow

From customer payment intent to cleared AR
Track A — high-volume, AI-enabled (target state)
01
Payment initiation
Customer pays via SCT Inst / open banking / PSP with structured reference
02
Bank-feed ingest
CAMT.054 intraday statement → ERP / TMS within minutes
03
Remittance capture
Portal / EDI 820 / IDP extraction → structured invoice list
04
AI matching
ML engine matches payment to invoices with confidence score ≥ threshold
05
Auto-post
SAP F-28 / Oracle AR auto-clears open items → DSO impact same day
06
Exception queue
Low-confidence items → human review with AI suggestion pre-loaded
Track B — SME / manual (current state for the long tail)
01
Payment initiation
Customer initiates bank transfer — reference incomplete or missing
02
Bank statement
MT940 daily download or manual portal extraction — T+1 at earliest
03
Remittance hunt
AR analyst searches email, calls customer to obtain invoice breakdown
04
Manual match
Analyst keys the match in ERP — error-prone, T+2 to T+5 lag
05
Post & clear
Manual F-28 post — DSO impact delayed by days or weeks
06
Unapplied cash
Unmatched amounts sit in suspense — distort AR ageing and DSO
05

Benchmark KPIs

World-class vs market average
>95%
STP rate · world-class
Market avg: 55–70%
<4h
Payment-to-post cycle · best-in-class
Market avg: 1–3 days
<0.5%
Unapplied cash · % of revenue
Poor: >3% — DSO distortion risk
>90%
Remittance capture rate
Before manual chase
<3%
Invalid deduction rate
World-class FMCG/retail
1 FTE
per €80–120M managed AR
AI-enabled target state
06

Method comparison matrix

Implementation effort vs STP impact
MethodTypical STP rateImpl. effortCost levelLatencyBest-fit volumeKey risk
Virtual IBAN + structured ref90–98%Low–MidLowIntradayAnyCustomer reference discipline
AI cash application (HighRadius / Billtrust)85–97%Mid–HighHighNear real-timeHigh / complexTraining-data quality
Customer self-service portal85–95%HighMid–HighReal-timeMid / highCustomer adoption
EDI 820 / EDIFACT REMADV95–99%HighMidBatch/real-timeHigh / complexPartner onboarding effort
CAMT.054 intraday bank feed80–92%Low–MidLow<1 hourMid / highReference quality
IDP / AI remittance extraction70–88%MidMidMinutes–hoursMid / highModel accuracy on edge cases
PSP webhook + reconciliation85–95%MidMid (+ txn fee)Real-timeLow–high B2CPSP settlement reconciliation
Open banking / A2A95%+MidLow (no interchange)Real-timeLow–highGeographic coverage gaps
RPA bot (legacy bridge)45–65%Low–MidLow–MidBatchAny (tactical)Brittleness, maintenance
Manual (email remittance + keying)0% STPNone (technical)High (FTE)1–5 daysLow / SMEScale, errors, DSO impact
07

Systems & vendors landscape

Key platforms by category
Vendor / platformTypePrimary functionSAP integrationVolume fitNotes
SAP S/4HANA (FSCM / EBS)ERPNative AR, EBS, cash application, FSCM collectionsNativeMid–complexF-28/F-32 posting, CAMT.053/054 standard, AI cash-app module in FSCM 2024
HighRadiusAIAI cash application, deductions, collections, e-invoicingSAP certifiedHigh–complexBest-in-class STP rates; native IDP remittance; strong FMCG/retail deduction module
BilltrustAICash application, credit, collections, payments networkSAP connectorMid–highBusiness Payments Network (BPN) — pre-matched remittance from 700k+ buyers
EskerAIO2C suite: billing, cash app, collections, remittanceSAP certifiedMid–highStrong in EU/FR; good IDP remittance extraction; mid-market sweet spot
AdyenPSPGlobal acquiring, local payment methods, reconciliationSAP connectorHigh–complexUnified commerce; strong reconciliation API; preferred for enterprise B2C/digital
StripePSPGlobal card + A2A, billing, financial-ops APIVia APILow–highBest developer experience; Revenue Recognition module; growing B2B invoice pay
KyribaTMSTreasury, bank connectivity, cash mgmt, payment factorySAP TRM bridgeHigh–complexSWIFT connectivity; multi-bank aggregation; payment hub for complex multi-entity
Tink / TrueLayerOpen bankingA2A payment initiation, bank data, VRPVia APILow–highPSD2 PISP; iDEAL 2.0 routing; real-time A2A with instant confirmation
Rossum / ABBYY VantageIDPIntelligent document processing for remittance PDFsVia APIMid–highLow-code training; confidence thresholds; integrates to any ERP via REST
UiPath / Automation AnywhereRPABank-portal scraping, email remittance, ERP keyingSAP GUI / APIMid (tactical)Tactical bridge; not a strategic replacement for API/EDI; interim only
Seeburger / OpenTextEDI/B2BEDI 820, EDIFACT REMADV, AS2 messagingSAP certifiedHigh–complexPre-mapped partner libraries; global network; complex deduction-code handling
08

Transformation roadmap

Phased approach from manual to AI-driven
Principle: don't boil the ocean. Start where the pain is highest (unapplied-cash backlog, largest customers without structured remittance), demonstrate STP improvement in 90 days, then extend. Technology follows process design — not the other way around.
Phase 1 · 0–90 days

Foundation & quick wins

Deploy virtual IBANs or structured creditor references for your top 20 customers. Implement a CAMT.054 intraday feed. Set up SAP EBS rules for auto-clearing on exact match. Target: 60–70% STP from current baseline.

Virtual IBANCAMT.054SAP EBS
Phase 2 · 90–180 days

Remittance automation

Deploy IDP (Rossum / HighRadius) for email-remittance extraction. Onboard a customer portal for mid-tier customers. Connect top EDI partners with REMADV. Target: 80–88% STP.

IDPPortalEDI 820
Phase 3 · 6–12 months

AI cash application

Implement an AI matching engine (HighRadius or SAP FSCM AI module). Train on 12–24 months of history. Integrate deduction management for short-pay resolution. Target: 90–95%+ STP.

AI matchingDeductionsFSCM
Phase 4 · 12–24 months

Predict & optimise

Payment-behaviour prediction for collections prioritisation. A2A / open-banking rollout for new customers. Treasury integration with AR for real-time cash forecasting. Target: 95%+ STP, DSO reduction 5–10 days.

Predictive AROpen BankingTMS
▤ Get in touch

Stuck below 95% STP?

We diagnose where your cash application leaks and build the phased roadmap to a touchless payment stack. A 30-minute call, no deck, no pitch.