SAS® Detection and Investigation for Insurance Features
- Provides an insurance-specific fraud data model.
- Consolidates historical data from internal and external sources – claims systems, watch lists, third parties, unstructured text, etc.
- Eliminates or reduces redundant or inconsistent data with the solution’s built-in data quality tools.
- Seamlessly integrates with existing third-party systems.
Advanced analytics with embedded AI and machine learning
- Provides a broad set of modern statistical, machine learning, deep learning and text analytics algorithms from within a single environment
- Enables you to improve fraud models by testing different approaches in a single run, and comparing results of multiple supervised learning algorithms with standardized tests.
- Provides an array of analytical capabilities, including clustering, different types of regression, random forests, gradient boosting models, support vector machines, natural language processing, topic detection and more.
- Continuously updates and improves models based on prior output results.
Rule and analytic model management
- Provides prepackaged heuristic rules, anomaly detection and predictive models, so you can harness the power of advanced analytics right out of the box.
- Lets you create and logically manage business rules, analytic models, alerts and watch lists.
- Enables you to customize analytical models to identify claims fraud not found by existing business rules.
- Enables easy management of the deployment, aggregation, scheduling, suppression and routing of simple or complex rules across multiple factors, such as parties, data sources and business lines.
- Lets you run groups of rules and models alone, in parallel or at different times (intraday, daily, weekly, monthly, etc.).
- Facilitates collaboration with other business units on model development.
Detection and alert generation
- Calculates the propensity for fraud at first submission, then rescores claims at each processing stage as new claims data is captured.
- Reviews claims early in the adjudication process so you can stop suspicious activity at the prepayment stage.
- Enables you to incorporate fraud detection methods into the process at the most appropriate points – e.g., cases where anomaly detection scenarios may require data that is not available until later in the adjudication process.
- Combines alerts from multiple monitoring systems, associates them with common individuals and provides a more complete perspective on the risk of particular individuals or groups.
- Prioritizes the investigative order of alerts by scoring them in real time, based on specific characteristics.
- Automatically routes alerts to appropriate team members based on user-set rules and requirements.
- Displays all evidence for each case on a dashboard that you can customize to accommodate your investigative unit's processes.
Social network analysis
- Provides a unique network visualization interface that lets you analyze related activities and relationships at a network dimension, and identify linkages among seemingly unrelated claims.
- Enables you to produce complete dossiers of networks surrounding a case, and gain fast access to full details on all related parties and networks.
- Produces independent and combined fraud scores, so you can assess overall risk on a customer, claim or network basis.
- Increases investigator effectiveness by enabling investigators to merge and delete network entities, and add annotations (text and images) to specific entities in a network.
- Provides time slider functionality, which enables you to see how activity in a network develops over a time horizon.
Search and discovery
- Enables free-text, field-based or geospatial searches across all data (internal and external).
- Lets you refine searches using interactive filters and facets that are customized for the claims and SIU teams.
- Provides full entity descriptions that include other linked entities, which you can open and explore to evaluate the likelihood of fraud.
- Provides an intuitive interface that lets you construct complex queries without the need to understand specific syntax. For example, you can use fuzzy searching, proximity searching and field boosting while restricting searches to specific entity types, fields, comments or insights.
- Systematically facilitates investigations using a configurable workflow.
- Stores all information pertinent to a case, including detailed investigation information – e.g., interview notes and evidence for criminal or civil prosecution, restitution and collections.
- Assesses overall fraud exposure, including losses due to fraud as well as fraud detected or prevented.
Flexible deployment options and analytical services
- Enables faster implementation (and faster ROI) when installed and administered at the SAS hosting site, eliminating the need for staff to oversee the system.
- Can be hosted at your site, with SAS providing implementation assistance and training.
- Can be fully integrated with your existing operations environment, workflow solution and business process management objectives, including thorough business process discovery and review to ensure your objectives are met or exceeded.