Always-On Features & Profiling

Generate stateless features of simple expressions and derived values and stateful features of aggregations in context that build over time.

• Efficiently transform raw data into decision-ready information about behavioral patterns and trends through transaction-level aggregation.
• Serve data features consistently and universally across your digital landscape to decision trees, tables, business rules, ML models, traditional analytics, business processes, case management, and all composable assets. 

Define and leverage calculated attributes that provide more relevant signals and context-aware insights. Discover signals and context-aware insights by defining and leveraging calculations such as “days since last purchase” or “number of transactions over a period of time.” 

Build real-time recursive behavioral transactional profiles with stateful, time series, and recursive feature execution.

• Build real-time behavioral context with transaction data.
• Summarize and profile any entity with complex calculations and time-series aggregations, across accounts and devices to better understand customer behaviors and preferences.  

Label, register, share, operationalize, and manage features as reusable assets. 

• Ensure alignment between your training and production data and feature sets, helping you get more predictive analytic and machine learning-enabled projects into production faster.
• Foster collaboration by sharing features between business analysts and data scientists — this promotes adoption and reuse across a wider range of enterprise services beyond just predictive models, such as decision logic, business workflows, and even customer communications.