Sydney, Australia | 9 December 2024

Rich Data Co (RDC) is pleased to announce the general availability of our AI Decisioning Platform V5.0. This latest release enables customers to accelerate their data-driven decision-making processes with two optional generative AI (GenAI) assistants – Genesis Assistant and Portfolio Assistant – and includes enhancements to Data Foundation, Genesis Studio and the Decision Workbench.

Key New Features and Improvements

1. Genesis Assistant

The Genesis Assistant is designed to help data science teams build and deploy AI models more efficiently. This intelligent assistant responds to complex technical queries using a comprehensive knowledge base – covering platform documentation, AI and data science best practices, and Python code generation.

Customers can also upload their own documents to tailor the assistant’s knowledge base. Leveraging AWS Bedrock’s Retrieval Augmented Generation and the Claude Sonnet model, this optional component can be activated at your discretion.

 2. Portfolio Assistant

Designed to empower portfolio managers and analysts, Portfolio Assistant simplifies natural language queries about lending portfolios. It provides quick insights into performance, risk exposures, growth opportunities, and more – without requiring deep analytical expertise. Built on RDC’s Teachable and Explainable framework, the assistant ensures transparency, interpretability, and alignment with your organisation’s objectives. Utilising AWS Bedrock for infrastructure and enterprise security, Portfolio Assistant is ready for production once onboarding is complete and any necessary governance approvals are complete. RDC can support customers through this process.

3. Enhancements to Genesis Studio

We have expanded Genesis Studio with pre-installed tools, including a CSV Editor and PDF viewer, making it easier to work directly with common file formats. Additional improvements include:

• The ability to download multiple files for local analysis and documentation.
• Full localisation support, consistent with the entire RDC platform.
• Access to Data Sensor and Feature Foundation databases for auditing and analysis.
• Improved usability of pre-packaged Python virtual environments, with enhanced commands to review environment details and requirements.

4. Data Foundation Enhancements

Customers now have greater flexibility with Data Sensors. Our configurable Data Sensor V1 adapts to custom account, customer, and transaction schemas, enabling faster setups and adjustments. A new “helper” function assists data engineers in configuring Data Sensors. We have also enhanced data profiling services:

• The File Scan Engine now automatically detects and analyses DATE data types, in addition to numeric and string types.
• The Data Value Scan Engine now provides minimum and maximum values for DATE data types and includes predefined special value checks.

 5. Batch Decision Strategy Execution Improvements

To improve efficiency and reliability, batch decision strategy execution now consists of separate services for generating and sending request messages. This approach ensures smoother workflows, better error handling, and automated retries in the event of failures.

6. Decision Workbench Enhancements

Version 5.0 introduces several enhancements to the Decision Workbench, making it easier to manage and optimise decision strategies.

• New PRE-ASSESSMENT Strategy Type: Identify suitable leads for pre-assessed offers, such as merchant lending prospects.
• Expanded Review Recommendation Strategy: Extend functionality for Continuous Credit Monitoring (CCM) without affecting existing users.
• Enriched Data Attributes: Classify data attributes as ‘input’ or ‘output’ and apply mathematical functions (sum, average, min, max, count) to array-type attributes.
• Past Decision Configuration: Leverage Feature Foundation to integrate historical decision outcomes directly into current strategies.
• Reason Code Configuration: Assign reason codes to both met and unmet conditions for added clarity in decision-making.

In addition to these features, we have introduced several bug fixes to improve stability and performance.

Looking Ahead

RDC is committed to supporting customers as they capitalise on the power of data-driven decisioning. For more information about these enhancements or assistance with implementation, please contact us.