Midships has adopted a different approach to APP Fraud Prevention
01
Targeted Approach
Our system is designed to proactively detect and prevent Authorized Push Payment (APP) fraud only. It is not meant to substitute other fraud detection methods for scenarios such as Account Takeover, and therefore, does not require an SDK.
02
Transparent & Deterministic
We utilize a white box approach, providing transparency and enabling thorough investigations with conclusive outcomes.
03
Real-Time Evaluation
Our online solution is built for real-time transaction analysis, ensuring prompt detection and intervention.
04
Customer Data Security
The Midships solution runs entirely within your secure environment, guaranteeing the confidentiality of your financial data.
05
Proven with real transactional data
Our strategic partnership with Malaysia's Financial Authorities, enables us to prove and refine our solution using real transaction data.
06
Defence in Depth
Our system comprises three key components designed to enhance detection while reducing customer inconvenience: a Blacklist for blocking transactions linked to confirmed fraudsters; a Real-Time Fraud Prevention Engine; and an Offline Fraud Detection Engine
How Our Solution Works

Machine Learning
Our system employs machine learning to create behavioural profiles of all senders and recipients, irrespective of their banking affiliations. These are used to detect anomalies that may suggest a sender is susceptible to fraud or a recipient is displaying patterns akin to known money mules and fraudsters. Machine learning enables our profiles to evolve continuously as the behaviours of senders and recipients shift over time.

Fraudster & Victim Interaction Analysis
Fraudsters typically target multiple individuals. While preventing every fraud attempt may be unfeasible, our goal is to identify a fraudster across multiple incidents. Our system assigns a score to each exchange between a sender (potential victim) and a recipient (potential fraudster). This score then influences all future interactions involving the sender and recipient with others, with the objective of detecting a fraudster within three to five transactions.

Mapping Relationships
We map the relationships between accounts to identify money mule syndicates and potential victims. This allows us to take proactive measures to mitigate further fraud.
Once a fraudster or money mule is identified, immediate action (via dynamic update of the blacklist) can be taken to block any further transactions. With proper interbank cooperation agreements, the temporary blocking of money mule accounts could facilitate easier recovery of funds for victims.