Loan Fraud Manager
How a Dutch mortgage provider replaced periodic sampling with real-time monitoring of every mortgage application, live in three weeks.
A mortgage book worth watching closely.
Our client is a Dutch bank with around €25 billion in assets, running its mortgage application process on Salesforce alongside a set of legacy systems.
Applications move through a long happy path — receipt, document collection, review, preliminary and final offer, notary and transfer, loan activation — with plenty of room, at every step, for the wrong document, the wrong actor or the wrong exception to slip through unnoticed.
Sampling only proves part of the book is clean.
During internal audits, the team spent considerable time trying to verify that controls actually prevented fraudulent behaviour. Because those audits relied on periodic sampling, they left gaps in coverage and left undetected fraud a real risk.
Fraud checks were still largely manual and not data-driven, so human capacity could not keep pace with a growing volume of applications. And because every release of the business-rule framework could quietly introduce a new risk, problems often went unnoticed until the next audit caught up with them.
One rule engine, watching every application.
DataLane implemented the Celonis Loan Fraud Manager, an app that combines process mining, data analysis and a rule engine to replace sampling with a fully data-driven approach that monitors 100% of the dataset instead of a slice of it.
Together with the bank, DataLane defined workflow, validation and document-validation rules tailored to its risk profile. Those rules are evaluated continuously, so a violation is flagged the moment it happens rather than at the next audit cycle.
Real-time dashboards give the fraud and audit teams a single, centralised view of the mortgage process, highlighting exactly when and where a rule was broken. Every violation is logged with a timestamp, building an audit-ready trail without extra manual work.
- Client
- Dutch bank, €25 billion in assets
- Process
- Mortgage application, origination to repayment
- Source systems
- Salesforce and legacy systems
- Platform
- Celonis Loan Fraud Manager
What made it hard to fix
- Fraud checks were performed by people, on a random, non-representative sample
- No real-time detection of control breaches or suspicious behaviour
- Every release of the business-rule framework could introduce fresh, unmonitored risk
- Strict regulatory requirements around segregation of duties and role-based access
What it delivered, and what made it stick.
Business value
- €750K in operational efficiency, from replacing manual checks and quarterly audit spend with automated monitoring.
- 100% process adherence on the controls in scope, instead of a sample.
- Full transparency for auditors, with a digitised, audit-ready control framework aligned to regulatory and internal standards.
- Fraud patterns and exceptions caught immediately, instead of surfacing weeks later in an audit.
Adoption
- Live in three weeks, on top of data already flowing from Salesforce and the bank’s legacy systems.
- Fraud and audit teams manage their own rules through a self-service portal, without needing to go back to IT.
- Automated alerts and case workflows route violations to the right team the moment they are detected.
- Version control on the rule framework flags any inconsistency introduced by a system update before it becomes a blind spot.
What it looks like day to day.
Recognise this in your own process?
Tell us where orders, cases or invoices go missing. We’ll show you what the data says.
Start a conversation