Discover Hyperlogy Fraud Detection, a solution that combines a rules-based approach with artificial intelligence (AI) to help banks detect unusual transactions, assign risk scores, and proactively prevent fraud in real time.
As Payments Get Faster, Fraud Prevention Must Become More Proactive
As real-time payments continue to expand, online scams and payment fraud schemes are becoming increasingly sophisticated. The speed at which funds move also makes fraud monitoring and prevention more challenging: if a suspicious transaction is detected only after funds have been transferred, investigating the transaction and recovering the funds may be more difficult.
Traditional controls often rely on lists of accounts flagged for suspected fraud or on post-transaction reviews. These measures may not be sufficient to identify new fraud schemes. In particular, schemes involving multiple intermediary accounts, funds routed through several layers, or individual transactions that appear legitimate require banks to analyze both the transaction context and the relationships between transactions.
This creates an urgent need for banks and financial institutions: they cannot rely solely on centralized alert lists, which may flag accounts only after they have been reported. They must proactively detect and prevent fraud within their own transaction flows.
Recognizing the critical nature of this challenge, Hyperlogy offers Hyperlogy Fraud Detection, an intelligent solution that helps banks detect and assess risk in real time within the transaction flow—before a transfer or payment instruction is sent to an external system.
>> Read more: Hyperlogy Successfully Supports NAPAS in Upgrading the National Fast Payment System to NAPAS 2.0
Market Context and Regulatory Compliance Pressures on Banks
Banks in Vietnam face pressure on two fronts: increasingly complex fraud and tighter regulatory requirements:
- Increasingly complex fraud: Online scams cause losses of trillions of Vietnamese dong each year. Tactics such as opening fake accounts, routing pooled funds through multiple layers, and rapidly cycling funds can unfold in seconds.
- Stricter compliance requirements: A series of regulations—including Decision No. 2345/QD-NHNN on biometric authentication, Circular No. 50/2024/TT-NHNN, and Circular No. 77/2025/TT-NHNN—require banks to strengthen the security of their online banking services.
- SIMO integration: Participation in SIMO—a collaborative account-monitoring system coordinated with NAPAS to track accounts suspected of fraud—has helped block suspected fraudulent transactions worth trillions of Vietnamese dong. However, to comprehensively protect their brand reputation and customers’ assets, each bank needs a robust, automated in-house fraud detection solution that can identify risks early and proactively.
Hyperlogy Fraud Detection: Controlling Risk Within the Transaction Flow
Hyperlogy Fraud Detection helps banks detect, assess, and handle transactions showing signs of unusual activity in real time. The solution combines two complementary approaches: a rules-based business rules engine and artificial intelligence (AI). This enables banks to control known fraud scenarios while also identifying suspicious relationships and behaviors that may be difficult to detect with standalone rules.
The solution can be integrated into the bank’s transaction orchestration and connectivity layer, such as a Payment Hub, ESB, or API Gateway, and work alongside existing Core Banking systems and transaction channels. This allows transaction instructions to be assessed before they are sent to external payment systems such as NAPAS, CITAD, SWIFT, or payment intermediaries, depending on each bank’s architecture and integration scope.
1. Real-Time Transaction Assessment and Decision-Making
When a transaction instruction is received, the system collects the necessary information, such as the accounts involved, amount, time, transaction channel, and device. The data is checked against blacklists and whitelists, business rules, and risk scores generated by AI models.
Based on the assessment, the system can take actions in line with the bank’s policies:
- Allow: The transaction meets the control criteria and proceeds to processing.
- Alert: The transaction is flagged as potentially unusual for the risk management team to review.
- Hold for verification: The instruction is held pending approval or additional authentication, if configured.
- Block: The transaction is declined before the instruction is sent if its risk level meets the bank’s blocking criteria.
This enables banks to choose how to handle each transaction according to its risk level, rather than applying the same control method to every transaction.
2. Hyperlogy Fraud Detection Combines Rules-Based Controls and AI to Identify Multiple Types of Fraud
Rules-based controls: Managing known fraud scenarios
The rules engine allows banks to define control conditions based on transaction limits, frequency, time windows, beneficiary accounts, transaction channels, and other business criteria. Banks can also manage blacklists and whitelists and set different thresholds for customer segments or transaction channels.
Through the administration portal, business teams can configure rules without coding each change. The two-step Maker–Checker approval workflow, configuration versioning, and audit logs strengthen control and traceability and support the review and justification of decisions.
AI – Analyzing behavior and connections between accounts
While rules-based controls handle predefined conditions, AI modules add the ability to analyze more complex patterns of unusual activity, including:
- Identifying transactions that deviate from a customer’s usual behavior and behavioral profile.
- Analyzing account networks to identify intermediary accounts, suspicious fund aggregation and distribution, or circular transfers.
- Assigning risk scores to transactions and accounts based on the data provided.
- Suggesting new rules based on fraud patterns that have been detected and verified.
Combining these two approaches expands detection capabilities while maintaining clear oversight of transaction handling decisions.
3. Network Analysis to Detect Suspicious Connections
An individual transaction may not show enough signs of fraud on its own. When viewed as part of the network of connections among multiple accounts, however, unusual patterns may become easier to identify.
Hyperlogy Fraud Detection’s network analysis module helps identify patterns such as one account receiving funds from multiple other accounts within a short period, funds being passed through several intermediary layers, funds circulating between accounts, or groups of accounts with suspicious connections.
Combined with risk scores and business rules, account network information provides the risk management team with additional grounds to investigate, assess, and make appropriate decisions.
4. Empowering Risk Management Teams to Take a More Proactive Role
In addition to transaction processing, the solution provides an administration portal for configuring rules, monitoring operations, and managing cases that have been flagged or blocked.
Administrative functions include managing rules through a Maker–Checker workflow, monitoring blocked transactions, reviewing decision histories, managing configuration versions, recording audit logs, and generating operational reports.
This enables banks to adjust control policies more proactively to meet business requirements while maintaining traceability for reviews and decision justification.
5. Flexible Integration with a Focus on Service Continuity
Fraud Detection is designed to connect to the bank’s existing transaction integration layer. This allows it to screen transaction instructions from channels such as Mobile Banking, Internet Banking, corporate online banking, ATM/POS, QR payments, and other integrated channels, depending on the implementation scope.
The architecture can scale to handle varying transaction loads by running multiple processing instances in parallel. Timeout and incident-handling behavior can be configured according to the bank’s policies to help reduce the risk of service disruption.
The solution also supports on-premises deployment on the bank’s infrastructure, meeting requirements for data control and operations under the model agreed for the project.
Why Banks Choose Hyperlogy Fraud Detection
As a pioneering technology company with more than 23 years of experience supporting the financial and banking sectors, Hyperlogy offers Fraud Detection with several key advantages:
- High performance and low latency: Delivers decisions within tens of milliseconds, helping prevent service disruptions and a poor customer experience.
- Reliable failover: Automatically switches to a fallback mode according to policy if an incident occurs, ensuring bank transaction instructions do not get stuck in processing.
- A single integration point for all channels: Integrates through the Payment Hub/ESB to protect all existing and future transaction channels.
- On-premises deployment: Runs directly on the bank’s infrastructure to meet strict national data security requirements, with 24/7 support from engineers based in Vietnam.
- Proven Implementation Experience: Hyperlogy has successfully developed and implemented Fraud Detection capabilities as part of the NAPAS 2.0 national payment system upgrade, demonstrating its expertise in fraud detection and large-scale payment infrastructure. The company also has extensive experience delivering digital banking systems for leading banks in Vietnam.
Hyperlogy: Proven Expertise Supporting the Banking and Financial Services Sector
Founded in 2003, Hyperlogy Corporation has more than 23 years of experience supporting the financial and banking sector. The company continues to demonstrate its capabilities and vision through major projects for leading financial institutions and banks in Vietnam, including Vietnam Bank for Agriculture and Rural Development (Agribank), Bank for Investment and Development of Vietnam (BIDV), Military Commercial Joint Stock Bank (MB), An Binh Commercial Joint Stock Bank (ABBANK), Saigon–Hanoi Commercial Joint Stock Bank (SHB), Bao Viet Commercial Joint Stock Bank (BAOVIET Bank), and Vietnam National Payment Corporation (NAPAS), among many other reputable organizations.
Hyperlogy’s expertise is grounded in practical experience. The company has a team of highly skilled technology experts who understand banking operations in depth and can work with advanced technology solutions. Its solution ecosystem is highly scalable in both transaction volume and functionality, according to project requirements. Integration with the Smart Digital Bank multi-channel digital banking ecosystem allows solutions to be customized and adjusted flexibly to meet each bank’s specific business needs.
