Behavioural Analysis
The automated behavioural analysis of customers, taking into account their profiles, transactions, and relationships, enables effective, precise monitoring of customer behaviour.
Together with the ongoing profiling of customers (module "Dynamic Profiling"), automated behavioural analysis forms the basis for the risk-based approach of the international Financial Action Task Force (FATF). It supports financial intermediaries in meeting regulatory requirements and in effectively detecting, preventing, and combatting money laundering, fraud, and cyber attacks. In marketing and sales, behavioural analysis provides precise scores for a potential-oriented approach, such as in sales campaigns or in the dynamic control of content in web applications.
The analysis of customer behaviour is based on self-learning algorithms from the fields of artificial intelligence and evolutionary learning. Based on this, models and rules are automatically created and continuously optimised, validated, and calibrated.
These data processing and analysis methods include both supervised learning for automatic recognition of known behavioural patterns and unsupervised learning for the search and identification of currently unknown, suspicious behavioural patterns with potential for opportunities or risks.
The model templates provided within this module are sharpened with company-specific data depending on the initial situation. This results in optimised models and rules that fully take into account customer specifics and the business model of the respective company. The highly automated and optimised processes in the Behavioural Analysis module ensure maximised result quality and high efficiency.
Using the same model templates for the entire community of an industry (e.g. banks, insurance companies) leads to enormous synergies for individual users without neglecting, ignoring, or affecting their peculiarities and characteristics when creating specific models.
With this option, Prospero further expands the functionalities of the DetectX® analysis platform. Together with the rules created by experts in the Rule Designer, it creates a framework for comprehensive and effective detection of opportunities and risks for a 360° view of customers.
The main functionalities
Analysis and modelling of customer behaviour to timely and accurately identify opportunities and risks.
Dynamic creation of models, rules, and risk profiles.
Dynamic recognition of different - i.e., already known and/or unknown - behavioural patterns based on a pattern-based recognition approach. This approach is based on an application of combined, self-learning machine intelligence methods optimised for this task.
Use of self-learning algorithms (evolutionary learning, supervised learning, and unsupervised learning) as well as unique optimisation methods.
Calibration of user-optimised models (not generic or generally applicable modelling) based on application-specific model templates for different industries (e.g. banks, insurance), which fully take into account the individuality, particularity, and business model of the customer.
Automated and ongoing dynamic adaptation of relevant model factors and weights.
Automated and continuous optimisation, calibration, and validation of models.
Ongoing monitoring of the quality and stability of models.
Maximised reduction of false-positive and false-negative detections, maximised detection of true-positive and true-negative results.
System-generated recommendations for optimal error-cost ratios, taking into account user-specific features as well as current data quality and quantity, with the aim of minimising losses and costs and maximising profit and savings.
Complete model life cycle management with full transparency, traceability, and revision security for the modelling process, changes, adaptations, and versioning.
Interactive pending system with checks and escalations, comprehensive documentation, and individual configuration of processes within the interactive alert viewer, e.g. supervisor and compliance checks, decision-making authorities, queries, and validations.
Flexible creation and maintenance of additional lists, criteria, and reference tables with the integrated rule designer, e.g. FATF high-risk countries, criteria for transactions and risky business relationships (high-risk transactions, high-risk business relationships), thresholds and risk criteria for complex structures.
Added Value
Dynamic, timely and automated detection of opportunities and risks.
Detection of known, unknown and complex patterns of behaviour and their precursors, as well as hidden or non-linear relationships.
Augmentation of existing expert knowledge and business rules through the automated generation of dynamic rules and detection models.
Continuous optimisation, validation and calibration of created models through self-learning and self-improving processes.
Creation of customer-specific calibrated models based on pre-configured, task-specific base models (no generic models) taking into account the business model and customer specifics of the user.
Automated calibration and adjustment to changing conditions.
Maximised recognition quality.
Complies with legal requirements and can be adapted to changing requirements at any time.
Integration modules for easy integration into existing systems.