What data does AR automation read from and write to the ERP?

Key Takeaways

• AR automation reads six data sets from and around the ERP and writes back a short list of postings, each after a person approves it.
• Among users of cash application software assessed by The Hackett Group, the median auto-match rate is 70% (The Hackett Group, 2025).
• ClearMatch from Transformance reaches a 95% auto-match rate and runs end to end in under 8 minutes (Transformance customer data, 2026).
• The ERP connection typically takes 2-3 days (your IT team may be involved); the full rollout takes up to 4-8 weeks.

In This Article

• Key Takeaways
• What data does AR automation read from and write to the ERP?
• Which fields decide whether a payment matches and a posting goes through?
• Which data comes from outside the ERP?
• What does AR automation write back to the ERP?
• How does the field map change across company codes, subsidiaries and legal entities?
• How does Transformance connect to read and write this data?
• What does Transformance do with the data it reads?

What data does AR automation read from and write to the ERP?

AR automation reads customer master data, open items, payment history, bank statements, remittances and journal entries, and writes back clearing postings, journal entries and reversals after approval.

What AR automation reads in each ERP

What AR automation writes back in each ERP, after approval

Object names describe where each ERP keeps the data. The connection method and the exact objects follow your environment. Oracle is supported; the other ERPs in the table run in production with customers. Remittance emails and PDFs usually sit outside the ERP; journal entries are read where each ERP keeps them.

‍

‍

A field map is the list of ERP records and fields an AR tool reads, the records it writes, and the direction of each, agreed before the connection is built. Read wide, write after sign-off: reads change nothing in the ledger, and each write is a posting the ERP validates (tolerances, open period, document type) after a person has signed it off.

Why the map matters: among end users of cash application software assessed by The Hackett Group (2025), the median auto-match rate is 70%, and a third of them exceed 80%. In our experience, much of the unmatched rest is a data problem: a reference in an unmapped field, a payer the customer master does not know, a remittance that never reached the ERP.

The same action has different names: SAP clears, NetSuite, Business Central and Oracle apply, Dynamics 365 F&O settles, Odoo reconciles. Table-level names are in where open items sit in SAP, Oracle and NetSuite, FI-AR in depth in how AI agents work with SAP.

Which fields decide whether a payment matches and a posting goes through?

Six groups of fields decide a match and a posting: the payer’s identity, the open item’s references and amounts, the bank line, the remittance, payment history and the posting fields.

The fields that decide a match or a posting

Reference fields are the most common gap. Customers quote their own PO number, or your invoice number in their own format, so read the open item’s reference fields as well as the document number: reference and assignment in SAP, PO number and memo in NetSuite, External Document No. in Business Central. Agree in writing before go-live which field holds it; few mapping decisions change the match rate more. For payments that quote nothing, see how to match a payment with no reference.

A correct match can still fail at posting. Each ERP documents its own failure modes:

• SAP tolerances. SAP checks a payment difference against two tolerance groups, the employee’s and the customer’s: “The payment difference must be within both tolerances to be handled automatically”, and “the entries in the tolerance groups are always in the local currency” (SAP Learning, S/4HANA course).
• Dynamics 365 F&O marks. An invoice cannot be selected for settlement when it is already marked by “another customer payment journal, sales order, vendor payment journal, or purchase order in the current legal entity or another legal entity” (Microsoft Learn, 2024).

Which data comes from outside the ERP?

Remittance advices usually arrive outside the ERP, as emails, PDFs, portal downloads or EDI files, so AR automation reads them alongside ERP data and links them to open items.

The bank statement does reach the ERP, as MT940, camt.053 or BAI2. But the invoice list behind a lump-sum payment often sits in an email attachment, and the note to payee carries a reference or nothing. Each ERP has a slot for the reference, not for the remittance (the Remittances row above); in Oracle, remittance references feed the Apply Receipts Using AutoMatch process (Oracle, 26D, 2026).

So the read map has to include the inbox. DocSense from Transformance reads remittances in any format with over 97% extraction accuracy across 100+ document types and 170+ languages, with zero template configuration (Transformance platform data, 2026). Mechanics: how AI reads remittance advice.

Illustrative example (companies and figures are fictional). A customer pays EUR 30,150.00 with the note to payee “RE 90012 90015”. Three invoices are open: 90012 for EUR 12,000.00, 90015 for EUR 9,400.00 and 90021 for EUR 9,250.00, together EUR 30,650.00.

The bank line names two invoices, but 90012 and 90015 add up to only EUR 21,400.00. The PDF remittance explains the remaining EUR 8,750.00: it pays 90021, less a EUR 500.00 deduction for a damaged pallet.

• Reads: the bank line, the PDF, the customer master (the IBAN identifies the payer) and three open items.
• Writes, after approval: 90012 and 90015 cleared in full; 90021 cleared with EUR 500.00 left open. In SAP that is a residual item with a reason code, in Business Central a partial application that leaves EUR 500.00 open on the ledger entry.

What does AR automation write back to the ERP?

AR automation writes clearing postings, payment differences, unapplied cash, journal entries for differences and charges, and reversals; each write follows your approval thresholds and passes the ERP’s own validation.

Five kinds of write, in each ERP’s terms:

1. Clearing. In SAP, “a clearing transaction always creates a clearing document”, and each cleared item records the clearing document number and clearing date (SAP Learning, S/4HANA course). In NetSuite, a customer payment transaction “records a customer payment and applies it to the appropriate invoice or cash sale, decreasing the amount due” (Oracle NetSuite, help center); NetSuite and Oracle in more depth: how AI agents work with Oracle and NetSuite. In Odoo, an invoice moves from “In payment” to Paid once “reconciled with a bank transaction” (Odoo, 18.0 docs).
2. Payment differences. In SAP, a partial payment means “all the documents remain in the account as open items”; a residual item means “the original document and the payment are cleared” and only the residual item remains, with reason codes to “describe the reason for the payment difference” (SAP Learning). Dynamics 365 F&O and Business Central settle or apply a partial amount and leave the rest open. Your accounting policy picks the route; the full flow is in posting payments and clearing open items, deductions in automating partial payments and deductions.
3. Unapplied cash. When the invoices are unknown, the payment still lands on the customer (the Unapplied cash row above). Dynamics 365 F&O says it plainly: “You’re not required to settle the payment to any invoices” (Microsoft Learn, 2025). In NetSuite, unapplied payments can later be applied to invoices on the customer payment page (Oracle NetSuite, help center).
4. Journal entries. Small differences, bank charges, currency differences and deduction postings, as your policy allows. ClearMatch applies 100% posting validation and nothing posts without human sign-off; the ERP then runs its own checks (balanced debits and credits, an open period, valid accounts).
5. Reversals. Transformance can post reversals back to the ERP, and each ERP undoes a clearing its own way (the Reversals row above). In Business Central, Unapply Entries creates correcting entries that “are the same as the originals but have an opposite sign”, and “the entries that the application closed are reopened” (Microsoft Learn, 2025). The audit side: what an AI agent’s audit trail records.

Thresholds by amount, entity and customer decide who approves each write. Admins configure the rules, finance approves, and every ERP posting sits inside the 4-level approval model (read, recommend, execute, post), detailed in who approves what when an AI agent works in your ERP. Other ERP updates, such as a note on the customer record, are agreed per project.

How does the field map change across company codes, subsidiaries and legal entities?

Every read and write carries the entity: SAP company code, NetSuite subsidiary, Dynamics 365 legal entity, Business Central or Odoo company, Oracle business unit.

Transformance consolidates multiple SAP instances and company codes, NetSuite subsidiaries and Dynamics 365 legal entities in one environment. Reads see the whole group; each write goes back to the entity that owns the open item, in that entity’s period and currency.

Three situations change the map:

• One payer, two entities. A customer pays invoices of two of your entities in one transfer. Writing needs one posting per entity; whether cash moves through an intercompany clearing account is your policy. Group view: multi-entity order-to-cash.
• Marks across entities. In Dynamics 365 F&O, an invoice marked for settlement in another legal entity cannot be selected (Microsoft Learn, 2024), so a cross-entity view has to show the mark before anyone approves.
• Currencies. SAP keeps tolerance groups in local currency, so read the transaction and the local currency of each item.

How does Transformance connect to read and write this data?

Transformance connects through the method your environment supports (APIs, SAP RFC/BAPI or IDoc, file/SFTP, middleware or its on-premises agent), typically in 2-3 days; your IT team may be involved.

SAP Cloud Connector and a customer Azure app registration are options too. For SAP, the connection keeps a clean core: no custom ABAP and no modifications in your SAP system.

The connection is one step of the rollout: the full rollout takes up to 4-8 weeks, and a proof of value runs in 4-5 weeks on one entity (implementation time).

‍

‍

• Access: your people sign in through SSO/SAML, and role-based access control (RBAC) sets what each person sees and approves. Security review: how AI agents access ERP data safely.
• Data: hosted in your region, with VPC deployment in your own cloud available (data in your region). No customer data is used to train models.

Per-ERP depth sits on the integration pages: SAP S/4HANA, SAP ECC, NetSuite, Dynamics 365 Finance & Operations, Business Central, Odoo and Oracle (supported). The overview: which ERP systems Transformance integrates with.

What does Transformance do with the data it reads?

Transformance keeps what it reads in Customer 360, one graph per customer; Vero, its agent, matches payments, investigates exceptions with evidence and proposes each write-back for a person to approve.

ClearMatch reaches a 95% auto-match rate and runs end to end in under 8 minutes (Transformance customer data, 2026). Set next to the 70% median above, that is what complete field data and remittance reading can add.

For a payment the rules leave open, Vero starts from Customer 360: the payer’s history, group links, open items, past deductions. It fetches further internal and external information to find more matches, then proposes the match with the evidence attached. A person approves, changes or rejects it.

The audit log records the timestamp, what the system did, what the user did, who approved, the input and the confidence. The ERP stays the system of record: Transformance reads it, proposes, and writes back the postings a person has approved.

Who Transformance is for

• ERPs: finance teams on SAP S/4HANA or SAP ECC, NetSuite, Microsoft Dynamics 365 Finance & Operations or Business Central, or Odoo, all running in production with customers today; Oracle and Infor are supported, and several SAP instances, NetSuite subsidiaries or Dynamics 365 legal entities consolidate in one environment.
• Size: mid-market to large enterprise, 250 to 2,000+ employees, including groups with several entities and shared service centres.
• Payment volume: from under 1,000 to over 100,000 payments a month.
• Industries: manufacturing, consumer goods and food and beverage, logistics, e-commerce, agriculture, energy and professional services.
• Scope: cash application alone, or the full order-to-cash (cash application, deductions, AR collections, cash forecasting).

Conclusion: map the fields before you connect

Agree the field map per ERP and entity first: what is read, which few postings are written, and who approves them. Then the connection is a matter of days, and the match rate rests on fields you have already checked.

Key terms

• Sub-ledger: the detailed customer ledger behind the receivables balance in the general ledger.
• Journal entry: a dated posting with balanced debits and credits, the form every write-back takes.
• Remittance advice: the customer’s list of invoices and deductions a payment covers, often outside the ERP.
• On-account cash: a payment posted to the customer without clearing specific invoices.
• camt.053: the ISO 20022 end-of-day bank statement most ERPs import.
• Audit trail: the record of who did, proposed and approved each action.

Related guides

• Which ERP systems does Transformance integrate with?
• How do external systems post payments and clear open items?
• Who approves what when an AI agent works in your ERP?
• What does an AI agent’s audit trail in finance record?
• How does Transformance integrate with SAP S/4HANA?
• How do AI agents access your ERP and customer data safely?

Frequently Asked Questions

What data does AR automation need from the ERP?

AR automation needs customer master data, open items, payment history, bank statements and journal entries from the ERP, plus the remittances that arrive by email, PDF or portal.

Does AR automation only read from the ERP, or does it write back?

Transformance reads and writes: it writes back clearing postings, journal entries and reversals, each after a person approves it and after posting validation.

Which ERPs and editions does Transformance read from and write to?

SAP S/4HANA, SAP ECC, NetSuite including OneWorld, Microsoft Dynamics 365 Finance & Operations and Business Central, and Odoo run in production with customers today; Oracle and Infor are supported. For SAP S/4HANA, Transformance connects to on-premises, private cloud and public cloud.

How long does it take to connect AR automation to our ERP?

The ERP connection typically takes 2-3 days, and your IT team may be involved; the full rollout takes up to 4-8 weeks, and a proof of value runs in 4-5 weeks on one entity.

Can one environment read several company codes, subsidiaries or legal entities?

Yes: Transformance consolidates multiple SAP instances and company codes, NetSuite subsidiaries and Dynamics 365 legal entities in one environment, and posts back to the entity that owns each item.

Does it work at our payment volume?

Transformance is built for anything from under 1,000 to over 100,000 payments a month, with a 95% auto-match rate and an end-to-end run in under 8 minutes.

What does it cost to connect AR automation to the ERP?

The cost depends on scope (cash application alone or the full order-to-cash) and on the number of entities and ERPs. Book a Call for a quote on your setup.