Core Banking System
The system of record holding accounts, balances and transactions — and the constraint most banking interfaces are actually designed around.
The core is where accounts, balances, ledgers and transaction processing live. Everything customer-facing is a layer over it, and its characteristics leak straight through to the interface: whether balances update in real time or after an overnight batch, whether a transaction can be enriched with a merchant name, how long a transfer takes to appear, what can be changed after submission.
This is why banking UI problems are so often not UI problems. “Why is my balance wrong” is usually a core running end-of-day batch rather than a rendering bug. Designing against a legacy core means designing around latency honestly — showing pending and available balances as distinct, stating when something will settle — instead of pretending to a real-time experience the system cannot deliver.
In practice
A bank's app showed one balance figure from a core that settled overnight. Customers spent, saw an unchanged balance, spent again, and overdrew. Splitting the display into available and pending, with the settlement time stated, cut overdraft-related complaints by about 40% without any core change.
Where teams get it wrong
- Presenting a single balance when the core distinguishes available from pending.
- Promising real-time behaviour a batch-based core cannot deliver.
- Designing flows that assume a transaction can be amended after submission when the core says otherwise.
- Treating core latency as a bug to hide rather than a fact to communicate.
- Starting interface design before anyone has asked the core team what is possible.
( Learn more )
Related terms
All terms- NeobankA digital-only bank with no branches — sometimes holding its own licence, sometimes operating on a partner bank's.
- Embedded FinanceFinancial products delivered inside a non-financial product, at the moment they are needed, instead of sending the user to a bank.
- Transaction MonitoringAutomated screening of payments against rules and models to flag suspicious activity for human review.
- Data TableA grid of rows and columns for scanning, comparing and acting on records — the screen most B2B users spend their day inside.