Detect and Prevent Agent and Merchant Fraud
Mobile money agents and merchants sit at critical points in the transaction ecosystem. Commission abuse, fraudulent cash-in and cash-out, merchant manipulation, and unusual relationships between participants can be difficult to identify when transactions appear legitimate individually.
Effective detection requires monitoring agent and merchant behavior alongside transactions, accounts, devices, customers, locations and relationships.
Explore how Neural Technologies' fraud management and Mobile Money Protection solutions help mobile-money operators and fintechs identify potentially fraudulent agent and merchant activity.
Understanding Agent Fraud and Merchant Fraud
Agent and merchant fraud refers to potentially fraudulent, abusive or unauthorized activity involving agents, merchants, their accounts, their channels or the transactions they facilitate or process.
Agents and merchants perform different roles, but both can become important risk points within a financial ecosystem.
Agent Fraud Risks
Agent fraud involves potentially fraudulent or abusive activity associated with a mobile-money agent, agent account, agent channel or agent-assisted transaction.
Agent-related fraud can take several forms, including:
|
Agent Fraud Type |
What It Can Look Like |
|
False transactions |
Customer funds transferred without proper authorization |
|
Unfulfilled transactions |
Customer provides cash but wallet is not credited |
|
Transaction splitting |
One transaction divided into multiple transactions |
|
Commission abuse |
Artificial or unnecessary activity to increase commissions |
|
Unauthorized cash-out |
Funds withdrawn without legitimate customer authorization |
|
Customer manipulation |
Customer persuaded to perform or authorize a fraudulent action |
|
Agent impersonation |
Fraudster poses as a legitimate agent |
|
Collusion |
Agent works with customers, fraudsters, accounts or other agents |
|
Registration abuse |
Accounts created or registered improperly to obtain financial or commission benefits |
|
Mule-account activity |
Agent connected to accounts used to receive or move suspicious funds |
Merchant Fraud Risks
Merchant fraud involves potentially fraudulent or abusive activity associated with a merchant, merchant account or merchant payment activity.
It is particularly relevant to fintechs, payment providers and mobile-money operators that offer merchant-payment services.
|
Merchant Fraud Type |
What It Can Look Like |
|---|---|
|
Fake merchant activity |
A fraudulent or misrepresented merchant identity or business |
|
Transaction manipulation |
Artificial, unauthorized or manipulated payment activity |
|
Refund abuse |
Suspicious or manipulated refunds or reversals |
|
Customer-merchant coordinated fraud |
Customers and merchants coordinate activity to generate fraudulent or unauthorized value |
|
Merchant account takeover |
Unauthorized access to a legitimate merchant account |
|
Onboarding abuse |
False, incomplete or misleading merchant information |
|
Transaction laundering |
Merchant activity used to disguise prohibited or suspicious transactions |
|
Unusual merchant activity |
Payment behavior that differs significantly from the merchant's established profile |
Why Agent and Merchant Fraud Is Different
Agent and merchant fraud can be difficult to identify because the participant is part of the legitimate transaction infrastructure. The transaction may therefore have:
- A valid customer
- A valid agent
- A valid wallet
- A valid transaction
- A valid destination
Yet the overall activity may still be abusive or unauthorized.
Scenario example:
An agent may normally serve a stable group of customers. A sudden increase in activity involving several related accounts, repeated transactions near a commission threshold and unusual device or location activity could provide a different risk context.
Similarly, a merchant can process legitimate customer payments while its overall behavior changes significantly, for example, through unusual transaction concentration, refund activity or relationships with connected accounts.
A valid transaction does not necessarily mean valid behavior.
These patterns may not be visible when transactions are assessed individually. Monitoring the relationships between participants, accounts, devices and transactions can provide additional context for investigation.
Common Agent and Merchant Fraud Scenarios
Commission and transaction abuse
An agent may generate artificial or unnecessary transactions to increase commissions or exploit transaction structures.
Transaction splitting can be one indicator. A transaction that would normally occur once may instead appear as several smaller transactions, potentially creating additional commission or avoiding a transaction threshold.
The important signal is not simply the number of transactions. It is whether the transaction pattern is unusual for that particular agent.
Unauthorized cash-in and cash-out activity
Agents can handle physical cash and digital value, creating opportunities for unauthorized or unfulfilled transactions.
For example, a customer may provide cash for a wallet deposit, but the expected digital value is not delivered. Alternatively, funds may be withdrawn without legitimate customer authorization.
Monitoring agent behavior alongside transaction records and customer complaints can help identify recurring patterns.
Registration and identity abuse
Agents may participate in customer or account registration, making the agent channel an important point for KYC and identity controls.
Potential indicators include unusual registration volumes, repeated use of similar information, activity inconsistent with an agent's expected customer profile, or relationships between newly registered accounts and existing suspicious accounts.
Agent or merchant impersonation
Fraudsters may attempt to appear as legitimate agents or merchants to obtain customer funds or credentials.
Providers can reduce this risk through authentication, customer confirmation, appropriate onboarding controls and monitoring for unusual activity associated with participant accounts and devices.
Coordinated agent or merchant activity
Fraud can also involve relationships between an agent or merchant and other participants.
For example, an agent may repeatedly transact with a small group of related accounts, while those accounts share devices, beneficiaries or subsequent transaction patterns. A merchant may similarly show unusual payment and refund relationships with a concentrated group of customers.
These patterns may not be visible when transactions are assessed individually. Relationship intelligence and connected-entity analysis can provide additional context for investigation.
Agent and Merchant Fraud Prevention
Prevention starts before the first transaction.
During onboarding, providers can establish the participant's identity, business information, ownership, operating location, authorized services and expected activity. These controls create a baseline against which future behavior can be assessed.
However, onboarding alone is not enough.
Participant risk can change over time as transaction volumes increase, new customers are acquired, devices change, locations change or new relationships develop. Continuous monitoring allows providers to identify those changes and apply risk-based controls.
Depending on the business model, prevention can include transaction limits, customer confirmation, secure transaction processes, commission controls, participant monitoring, behavioral analytics, investigation workflows and targeted intervention.
Effective prevention combines appropriate controls with continuous detection and monitoring as participant behavior changes.
Agent and Merchant Fraud Protection with Neural Technologies
Neural Technologies' Mobile Money Protection approach connects participants, accounts, devices, transactions and behavioral information to help mobile-money operators and fintechs identify potentially suspicious activity and assess fraud risk across the wider ecosystem.
For mobile-money providers, this can support monitoring of agent activity such as transaction patterns, cash-in/cash-out behavior, commission activity and changes from established agent behavior.
For fintechs and payment providers, the same approach can support merchant monitoring, payment behavior analysis, account relationships and changes from established operating patterns.
Explore Neural Technologies' Mobile Money Protection: Fraud, Risk and Financial Crime Prevention
Speak to Neural Technologies about agent and merchant fraud protection, participant-risk monitoring and mobile-money fraud management.
Agent and Merchant Fraud: Frequently Asked Questions
Agent fraud refers to potentially unauthorized or inappropriate activity involving a mobile-money agent. Detection can consider transaction activity, participant behavior, customer relationships and other relevant signals.
Merchant fraud refers to potentially fraudulent or abusive activity involving a merchant, merchant account or merchant payment activity. Relevant indicators can include unusual transaction behavior, refund patterns, changes in operating activity and relationships with customers or other participants.
Providers can monitor agent transaction behavior, cash-in and cash-out activity, commission patterns, customer-agent relationships, device and location information, account relationships, and changes from the agent's established behavior.
Fintechs can monitor merchant transaction behavior, payment patterns, refunds, transaction concentration, account and device relationships, operating locations and changes from the merchant's established behavioral profile. Combining these signals can help identify activity that may warrant further investigation.
Transaction splitting occurs when a transaction is divided into multiple smaller transactions, potentially to exploit a commission or pricing structure. Monitoring repeated transactions involving the same customer, agent, or account can help identify potential splitting.
An unfulfilled transaction occurs when a customer gives cash to an agent for a service such as cash-in or bill payment, but the expected transaction is not completed or the customer does not receive the corresponding value.
Controls can include agent due diligence, transaction limits, customer confirmation, secure transaction processes, commission monitoring, continuous agent-behavior monitoring, customer education, and investigation of suspicious agent relationships.
AI and machine learning can analyze large volumes of agent, transaction, behavioral, account, and relationship data to identify anomalies, behavioral changes, and potentially suspicious patterns. These capabilities can complement rules-based fraud detection.
