SAS® Intelligent Decisioning Features
Enterprise scale decisioning
- Delivers more than 5,000 real-time transactions per second.
- Response times of 10 milliseconds per transaction.
- SAS data access engines simplify integration with a variety of third-party applications at the data level.
- Ability to define and reuse decision variables from a variety of sources, including CSV and data tables, as well as supporting complex structures with data grids.
- A graphical drag-and-drop interface lets you assemble business rules, custom code and models into complex decision flows, which minimizes the need to write deployment code that joins these pieces together.
- Ability to define decisions by browsing centralized data, model and business rule repositories and selecting from existing assets.
- Create, manage and test custom code within a decision flow to integrate with business application REST APIs, databases, web service calls and open source Python.
- Control decision orchestration by adding condition logic (i.e., IF-THEN-ELSE) and using outputs from any preceding rule or model.
- Enhanced rule list view provides compressed, easy-to-read rules that let you readily identify logic definitions.
- Built-in version control for entire decision flows simplifies testing and validation.
- Configure and view decision flows in linked view as well as referenced view for ease of understanding.
- Create complex decision flows for both batch and real-time environments, simplifying IT integration and acceptance testing, as well as operational deployment.
- Use SAS Studio to build custom data queries and pass data to decision nodes.
Business-user-centric rules management
- Integrated business rule management platform enables fast rule construction, testing, governance and integration within decision flows.
- Rule version management capabilities improve tracking and governance during deployment, including deep linking to business rules from decision flows.
- Create complex business logic quickly within decision flows, including on-the-fly term development.
- Provides freeform rule-logic creation with full access to sophisticated functions, both predefined and user defined.
- Incorporate data quality functions in business rules.
- Lookup table integration to execute lookup for rule-logic checks and rule actions.
- Lookup table management for table import and updates gives you the ability to create lookups from SAS Visual Analytics tables.
- Lookup tables can be activated and locked at user discretion to support proper usage of the most current lookup tables within business rules.
- Lock down or augment rule versions.
AI/ML augmented decisions
- Simplify model inclusion by seamlessly drilling from decision flow through to model repository in SAS Model Manager.*
- Supports all SAS models (including computer vision and text analytics) and models developed in open source frameworks such as Python and R.*
- Apply governance workflows to models through SAS Model Manager.*
- Directly include Python models in decisions via code node.
- Execute models natively, without the need for translation.
- Ease of creating test cases by bringing in data from a variety of sources with built-in data mapping tool.
- Perform scenario tests by interactively entering test values, including an expected result.
- An autogenerated validation test ensures that testing is run in a manner that resembles the chosen deployment destination.
- Use a common environment for disciplined testing, change management, auditing and validation.
- Reporting and user logs for audit history simplifies IT testing for applications that call operational analytics as web services.
- Register multiple input tables for use within SAS Intelligent Decisioning, including testing, publish target validation and simulation.
- Save rule tests, test suites and log details for documentation and reuse.
- Use explicit and detailed rule-fire analysis for testing, refinement and rule auditing documentation prior to operational deployment.
- Graphical tool to analyze and validate decision paths for both development and production.
- SAS macros to build BI reports with trace information persisted in production for audit.
- Built-in facility to configure and persist decision variables in database for ease of access for the purpose of decision improvement.
- Centralized management of all decision assets, including requirements, with role-based access.
- Reuse decision elements across teams.
- Compare different objects and versions to understand and track changes for merging branches and easier collaboration across teams.
- Custom functions and formulas can be shared across the organization via the expression builder.
- Generate PDF documentation for decisions, rule sets, lookup tables and treatment groups.
Performance monitoring automation
- Provide performance reports and notifications.*
- Automatically retrain analytical models when they decay.*
- Swap champion with challenger models based on thresholds.*
- Automate the complete end-to-end model management process.*
- Provide workflows and rules that govern model execution.*
- Perform champion/challenger model comparisons.
- Take advantage of decision analysis capabilities, including simulation options.
- Track and view the lineage of components to conduct an impact analysis of changes.
- Get more information about your customers through detailed response history.
- Strategy performance monitoring and reporting with SAS Visual Analytics.
Enhanced term management
- Automated rule-to-term mapping includes type and domains from existing data dictionaries and tables.
- Rename terms and choose what to include/exclude from the input and output the definition of a decision flow.
- Dynamically add new terms as needed to simplify term definition, data type, input/output designation and length.
Streamlined deployment to multiple destinations
- Real-time deployment (via REST API):
- Micro analytic web service (MAS) provides fast, scalable web service deployment.
- Easily move complete decision flows into IT web service testing environments and production deployment.
- Supporting analytical scoring as a service, MAS execution operates in a self-contained and portable standalone architecture (with a minimal footprint).
- Scalable and performant batch processing in the cloud:
- SAS Cloud Analytic Services (CAS) provides a cloud-native, highly performant, batch processing engine for analytics and decisions.
- In-database batch deployment
- Execute business rules and analytical model scoring without moving the data.
- Includes extended support for the following Hadoop environments: Cloudera, Hortonworks, MapR, Pivotal and BigInsights.
- Supports in-database rule execution for models, rules and decisions for Hadoop and Teradata.
- In-stream deployment:
- In-memory threaded kernel processing simplifies integration with transactional systems, as well as IoT or in-stream computing.
- Seamless, transparent read, write and update access to data, regardless of source or platform.
- Supports multiple loading options for moving refined data from SAS into third-party data stores.
- Reuse DBMS metadata for analytical purposes.
- Machine learning and AI suggestions: Scans data and makes intelligent transformation suggestions using machine learning and AI.
- Self-service interface: Generates code automatically from an intuitive, point-and-click interface so nontechnical users can profile, cleanse, blend and move data without specialized skills or training.
- Integration into analytics pipeline: Integrates prepared data into the analytics pipeline automatically, creating a seamless data discovery and data preparation user experience.
- Data lineage: Explore relationships between accessible data sources, data objects and jobs.
- Metadata access: Access physical metadata information like column names, data types, encoding, column count and row count to gain further insight into the data.
- Visually explore data, discover new patterns, and create and share smart visualizations and interactive reports through a single, self-service interface.
- Leverage augmented analytics and advanced capabilities to accelerate insights and find hidden stories in your data.
- Easily share insights across channels, such as web, mobile and Microsoft Office applications.
- Create and deploy custom conversational chatbots through an intuitive, low-code visual interface.
- Get text responses, access data, reports and visualizations and even apply analytics and AI through a natural language, conversational interface.
- Configure bots within the SAS environment for easier access to insights or connect to external services to roll them out to the world.
*When SAS Model Manager, SAS Visual Machine Learning or SAS Visual Data Science is licensed with SAS Intelligent Decisioning.