Oracle PeopleSoft and Plaid integration

What moves between them

Transactions and account balances flow from Plaid into Oracle PeopleSoft on a daily or real-time schedule, depending on configuration. ml-connector retrieves the latest transactions from Plaid for each linked bank account, maps them to corresponding bank accounts in PeopleSoft, and validates them against the reconciliation tables. Account balances are compared to PeopleSoft's recorded balances to flag discrepancies. If configured, ACH and wire transfers initiated in PeopleSoft can be routed through Plaid, with confirmation posted back to PeopleSoft. The flow is read-heavy on the Plaid side and read-mostly on PeopleSoft, since transactions are appended records that do not require updates.

How ml-connector handles it

ml-connector stores Plaid API credentials and access tokens encrypted, and obtains per-user tokens through Plaid Link before transaction retrieval begins. On the PeopleSoft side, it connects using HTTP Basic Auth or OAuth2 bearer token depending on the PeopleTools version, and accepts the full hostname and port per customer instance, since PeopleSoft is self-hosted with no shared base URL. Plaid webhooks can be registered to push transaction updates, but ml-connector also polls at a configurable schedule to ensure transactions are not missed. Transactions are matched to PeopleSoft bank accounts by account number and routing number, then posted to the reconciliation table with a reference to the original Plaid transaction ID to prevent duplicates. Plaid transfer descriptions are truncated to Plaid limits (15 characters for RTP, 10 for ACH), and transfer confirmation is verified via Plaid webhook signature using JWT with ES256 and a SHA-256 body hash before updating PeopleSoft status. Retries follow exponential backoff, and every transaction and transfer carries a full audit trail for compliance and debugging.

A real-world example

A mid-sized financial services company runs Oracle PeopleSoft on-premise for general ledger and accounts payable. They operate five bank accounts across three institutions and currently perform bank reconciliation by downloading transaction files from each bank portal, importing them manually into a spreadsheet, and matching them against PeopleSoft recorded transactions and cleared checks. Reconciliation takes two full days each month and reconcilers frequently miss transactions posted out of order or as batch entries. By connecting PeopleSoft to Plaid, transaction and balance data flows automatically each morning from all five bank accounts. Transactions are matched against PeopleSoft records and flagged for manual review only when a mismatch is detected, reducing reconciliation time to hours and catching errors earlier in the process.

What you can do

• •Retrieve bank account transactions and balances from Plaid daily or on-demand, and map them to PeopleSoft bank accounts and cash GL codes.
• •Validate transaction amounts and dates against PeopleSoft recorded entries and flag discrepancies for investigation.
• •Initiate ACH and wire transfers from PeopleSoft through Plaid, with confirmation and status updates posted back to PeopleSoft.
• •Authenticate PeopleSoft with HTTP Basic Auth or OAuth2 depending on PeopleTools version, and handle Plaid Link token exchange and encryption.
• •Store all transaction, balance, and transfer records with complete audit trails for compliance and forensic replay.

Questions

How does ml-connector obtain access to bank accounts in Plaid?

ml-connector uses Plaid Link, a three-step OAuth-like flow in which each user grants the application permission to view their bank accounts. The user logs in to their bank through Plaid Link, and ml-connector receives a public_token that is exchanged for a long-lived access_token and item identifier. The access_token is stored encrypted and used for all subsequent API calls to retrieve transactions and balances for those accounts.

Which direction do transactions flow between PeopleSoft and Plaid?

The primary flow is from Plaid into PeopleSoft. Bank transactions and balances retrieved from Plaid are matched to PeopleSoft bank accounts and reconciliation records. In the reverse direction, ACH and wire transfers can be initiated in PeopleSoft and routed through Plaid, with the transfer status posted back to PeopleSoft once confirmation is received from Plaid.

How does ml-connector prevent duplicate transactions?

Each transaction retrieved from Plaid is matched to a PeopleSoft recorded entry by amount, date, and account number. ml-connector stores the Plaid transaction ID alongside the matched PeopleSoft record, so if Plaid is polled again, the same Plaid transaction ID will not be re-imported. Transfer confirmations are similarly tracked by Plaid transfer ID to prevent double-posting.