Loading
Process Data by Using Document AI

You are here:

  1. Salesforce Help
  2. Docs
  3. Agentforce Revenue Management

Process Data by Using Document AI

Use Document AI to extract structured data from the ingested payment advice and payment proof documents. Document AI reads each document and writes the extracted values into an output data lake object (DLO), a structured table whose columns represent the fields that you want to capture. You define one output DLO for payment advice and one for payment proof, and Document AI populates them as it processes your documents. Edit or delete any fields or tables in a DLO as needed.

Required Editions

Available in: Lightning Experience
Available in: Enterprise, Unlimited, and Developer Editions with the Revenue Cloud Billing license. Contact your Salesforce account executive for more information.
User Permissions Needed
To process data by using Document AI:

Data Cloud Architect permission set

To configure these output DLOs, complete both procedures, first for payment advice, then for payment proof. For more information, see Document AI and Create a Document Schema Configuration Manually.

Configure Output DLOs for Payment Advice

  1. In the Data 360 app, click Process Content | Document AI | New.
  2. Select the payment advice UDMO as the source object.
  3. Select PDF and Image as the file types available for configuring the document schema.
  4. In DLO configuration, create two tables with columns that represent the payment-level and invoice-level details from the payment advice document.
    PAYMENT-LEVEL DETAILS ON PAYMENT ADVICE INVOICE-LEVEL DETAILS ON PAYMENT ADVICE
    Payment Advice Id Payment Advice Line Invoice Id
    Payer Name Payment Advice
    Payment Amount Invoice Number
    Payment Date Applied Amount
    Reference Number Gross Amount
    Source Document Deduction Amount
    Transaction Description Deduction Code
    Payment Method Created Date Time
    Processing Status Last Modified Date Time
    Account Invoice
  5. View and publish your document schema.
  6. Open the output DLOs to check the field-level mappings, and rebuild the schema if required.
  7. Use the Query Editor to view the records.

Configure Output DLO for Payment Proof

  1. In the Data 360 app, click Process Content | Document AI | New.
  2. Select the payment proof UDMO as the source object.
  3. Select PDF and Image as the file types available for configuring the document schema.
  4. In DLO configuration, create a table with columns that represent the payment proof details from the payment proof document.
    DETAILS ON PAYMENT PROOF
    Payment Proof Id
    Transaction Reference Number
    Payer Name
    Payment Amount
    Booking Date
    Value Date
    Statement Date
    Source Document Name
    Source Document URL
    Credit Debit Indicator
    Payment Description
    Processing Status
    Created Date Time
    Last Modified Date Time
    Proof Type
    Account
  5. View and publish your document schema.
  6. Open the output DLO to check the field-level mappings, and rebuild the schema if required.
  7. Use the Query Editor to view the records.
How Document AI Structures a Payment
How Document AI Structures a Payment

Consider a single payment from Acme Corp. Acme sends a remittance advice (a type of payment advice) to pay two open invoices with one $10,000 bank transfer, and Acme’s bank issues a statement that confirms the deposit. Document AI extracts data from both documents into their output DLOs. Because the two documents describe the payment differently, their schemas differ too.

Because one payment advice can pay several invoices, the payment advice output DLO uses two related tables: one for the payment-level details and one for the invoice-level records.

PAYMENT ADVICE DETAILS CONFIGURATION OF THE OUTPUT DLO
Payer Acme Corp, a $10,000 bank transfer dated March 15, 2026, with reference number ACME-RA-20260315. One record in the payment-level table: Payer Name, Payment Amount, Payment Date, Reference Number, and Payment Method.
Invoice INV-1001, with $6,000 applied and no deduction. One record in the invoice-level table: Payment Advice Invoice Number, Applied Amount, and Gross Amount.
Invoice INV-1002, with $4,000 applied against a $4,200 gross amount and a $200 early-payment deduction (code EARLYPAY). Another record in the invoice-level table, including Deduction Amount and Deduction Code.

A payment proof confirms a single payment that the bank received, so its output DLO uses one table, with one record per confirmed transaction.

PAYMENT PROOF DETAILS CONFIGURATION OF THE OUTPUT DLO
An incoming credit from Acme Corp for $10,000. Payer Name, Payment Amount, and Credit Debit Indicator.
Transaction reference TXN-88231, described as ACH CREDIT ACME CORP. Transaction Reference Number and Payment Description.
A booking date and value date of March 18, 2026, on a statement dated March 31, 2026. Booking Date, Value Date, and Statement Date.
The source file AcmeBank_Statement_Mar2026.pdf. Source Document Name and Source Document URL.

Together, these two output DLOs give payment reconciliation the structured data that it needs to later match Acme’s payment advice against the bank’s payment proof and the open invoices.

 
Caricamento
Salesforce Help | Article