The Platform takes in a diverse set of data sources to build a comprehensive understanding of each transaction. These data sources include transaction data, historical profile data, mobile device identifiers from our mobile SDK and web collection script, and data from our Global Data Network consortium, covering around 2000 attributes in addition to first- and third-party data sources. This data is aggregated, normalized and enriched using our proprietary data science within the Outseer Risk Engine.
Data Sources
Data Aggregation and Enrichment
Once the data is gathered, the Outseer platform aggregates the data inputs, enriching them to create a holistic view of each transaction. This step ensures that the data is comprehensive and ready for detailed analysis.
The enriched data is then processed by the Outseer Risk Engine. The Risk Engine utilizes our proprietary AI and ML approach to analyze the data and generate a Risk Score. The Risk Engine normalizes the raw score, which is correlated to the conditional probability of the transaction being fraudulent, to produce a score from 0 to 1000 according to a predetermined score distribution. The Risk Engine only uses the relevant data for the use case, so the score is tailored to every digital transaction. The benefits of this approach result in a model of high performance with a superior ability to explain the main indicators and factors driving your score.
The Platform then makes the data elements and the risk score available to the Outseer Policy Manager so you can render decisions based on specific risk profiles, customer segments and transaction types using different score decisioning thresholds, business rules and regulatory rules set by your financial institution. The Policy Manager also determines whether a user is challenged with step-up authentication, including what type of authentication.
Transactions identified as high-risk or potentially fraudulent, including any transaction that was not allowed by the Policy Manager, are escalated to the Outseer Case Manager. This application allows Fraud Managers to investigate and resolve these transactions. Once resolved, feedback is instantly provided to the Outseer Risk Engine and Global Data Network. The Risk Engine uses supervised machine learning algorithms that produce optimal results with feedback from Case Manager and authentication feedback. This continuous feedback loop ensures that our risk models are constantly learning from past events.
The Platform takes in a diverse set of data sources to build a comprehensive understanding of each transaction. These data sources include transaction data, historical profile data, mobile device identifiers from our mobile SDK and web collection script, and data from our Global Data Network consortium, covering around 2000 attributes in addition to first- and third-party data sources. This data is aggregated, normalized and enriched using our proprietary data science within the Outseer Risk Engine.
Data Sources
Data Aggregation and Enrichment
Once the data is gathered, the Outseer platform aggregates the data inputs, enriching them to create a holistic view of each transaction. This step ensures that the data is comprehensive and ready for detailed analysis.
The enriched data is then processed by the Outseer Risk Engine. The Risk Engine utilizes our proprietary AI and ML approach to analyze the data and generate a Risk Score. The Risk Engine normalizes the raw score, which is correlated to the conditional probability of the transaction being fraudulent, to produce a score from 0 to 1000 according to a predetermined score distribution. The Risk Engine only uses the relevant data for the use case, so the score is tailored to every digital transaction. The benefits of this approach result in a model of high performance with a superior ability to explain the main indicators and factors driving your score.
The Platform then makes the data elements and the risk score available to the Outseer Policy Manager so you can render decisions based on specific risk profiles, customer segments and transaction types using different score decisioning thresholds, business rules and regulatory rules set by your financial institution. The Policy Manager also determines whether a user is challenged with step-up authentication, including what type of authentication.
Transactions identified as high-risk or potentially fraudulent, including any transaction that was not allowed by the Policy Manager, are escalated to the Outseer Case Manager. This application allows Fraud Managers to investigate and resolve these transactions. Once resolved, feedback is instantly provided to the Outseer Risk Engine and Global Data Network. The Risk Engine uses supervised machine learning algorithms that produce optimal results with feedback from Case Manager and authentication feedback. This continuous feedback loop ensures that our risk models are constantly learning from past events.
An additional step-up authentication layer can be employed to further validate someone’s identity in high-risk scenarios or scenarios that violate your organization’s policies. This can include:
The Risk Engine a uses a probalistic model that modifies its future risk predictions based on fraud volumes and risky behaviors seen in association with the specific feature value. The scoring task recalculates and updates the weights for the different feature values, based on the ratio between the fraud distribution and the genuine distribution of each predictor seen in the most recent fraud and genuine evidence.
Dynamic training uses customer case markings to deliver a custom-trained model that targets your specific fraud to increase fraud detection while maintaining low intervention rates. It provides quicker detection of new sophisticated attacks, identifies fast moving fraud patterns, and provides more accurate models per use case.
Our KPI dashboard surfaces key metrics such as fraud detection rates, intervention rates, transaction volumes, and transaction value, allowing you to track performance, as well as benchmark against industry norms.
Outseer enables a secure and frictionless mobile experience. Our Mobile SDK is available for both iOS and Android devices. This SDK integrates with your mobile application, collects mobile device identifiers for risk assessment, and invokes biometrics and OTP push notifications as step-up authentication for flagged transactions.
