Rich Data Co (RDC) Enhances Credit Decisioning with AI-Driven Assistants Using Generative AI

RDC is a software-as-a-service (SaaS) provider dedicated to broadening access to sustainable credit through improved financial inclusion for commercial and business lending. Its SaaS platform helps banks and financial institutions make more informed, high-quality lending decisions by delivering deep insights into customer behavior. Despite advancements in AI, traditional credit assessment methods still rely on limited data, which can lead to evaluations of creditworthiness that are less comprehensive. RDC’s mission is to help banks leverage broader data sources, including transactional and behavioral information, to better assess customer viability and business health.

Gordon Campbell, co-founder and chief customer officer at RDC, explains, “There's more data in the world than ever before, with greater compute power and advancements in AI, yet traditional scoring models often overlook critical insights that advanced AI tools can capture.” Recognizing the need for AI-driven enhancements to make accurate decisions with vast amounts of customer data, RDC set out to develop two intelligent AI-driven “assistants” to support data scientists and portfolio managers. These assistants empower banks to build and analyze decision models, enabling high-quality lending decisions efficiently and securely.

RDC engaged AWS to integrate AI-driven assistants into its platform, addressing critical needs in model development, data querying, and portfolio analysis. RDC’s relationship with AWS began at the company’s founding when it chose AWS as its cloud provider. Over time, AWS became RDC’s trusted choice due to its established presence in the financial sector and its robust security and scalability. "The depth of knowledge at AWS really accelerated our progress, guiding us to overcome potential roadblocks and achieve our goals faster," explains Hendra Suryanto, chief data scientist at RDC.

To expedite development, RDC collaborated with AWS Startups and the AWS Generative AI Innovation Center to build two key AI-driven assistants on Amazon Bedrock: the Data Science Assistant and the Portfolio Assistant. Utilizing AWS-supported models and tools, these assistants help data scientists and portfolio managers conduct model development, troubleshoot queries, analyze data, and gain insights into various segments.

The Data Science Assistant is designed with tools to streamline model validation and provide advanced data insights. Using Anthropic Claude in Amazon Bedrock, it generates accurate responses to technical questions and leverages retrieval-augmented generation for quick access to a knowledge base stored in Amazon OpenSearch Serverless. Thanks to this setup, the assistant can pull relevant information on demand, boosting productivity for RDC’s data science teams.

The Portfolio Assistant converts text to SQL, helping portfolio managers perform data-driven inquiries into loan portfolios without needing deep SQL expertise. Structured portfolio data stored in Amazon Aurora MySQL is accessed through natural-language-to-SQL conversion, enabling the assistant to answer queries and provide insights on portfolio performance, risk exposure, and growth opportunities. The Portfolio Assistant also handles high-level portfolio questions, manages iterative interactions for complex queries, and powers self-correction tools including query validation and result accuracy checks, and prompts for additional information when needed.

Charles Guan, co-founder and CTO at RDC, says, “Amazon Bedrock is compelling not just because of the models but because of the overall environment as an ecosystem. The combination of Amazon Bedrock and guidance from AWS doubled our development speed and gave us confidence in meeting security and enterprise-grade requirements.” With this setup, RDC has eliminated infrastructure complexities and shifted its focus to creating high-impact, AI-driven assistants that align with its mission of financial inclusion and regulatory compliance.

With the AI-driven assistants developed on Amazon Bedrock, RDC has significantly enhanced its platform, facilitating more precise and efficient credit assessments in business lending. Through an iterative approach, RDC rapidly advanced its generative AI capabilities, deploying the initial version to production in just three months. The solution also meets the stringent security standards required in regulated banking environments. Currently in pilot with leading Australian banks, RDC’s Data Science Assistant is estimated to increase the volume of pre-assessed credit offers for existing customers from below 50 percent to well above 50 percent, giving more businesses access to credit with fewer barriers. The platform’s observability also supports early identification of financial distress, empowering institutions to proactively assist at-risk clients and promote healthier portfolios.

As RDC brings this new offering into early adoption cycles with select financial institutions, the company is actively collaborating with beta customers to refine the solution, aligning on security, explainability, and compliance. “Integrating generative AI on AWS into our platform marks a pivotal moment in our mission to transform credit decision-making,” Campbell says. “By equipping both data scientists and portfolio managers with AI-driven assistants, we're not just improving efficiency—we're redefining how financial institutions approach lending.”