Teachable Intelligence:

The art and science of AI Decisioning for business and commercial lending

Sydney, Australia | 28 February 2024

Business and commercial lending decisions often entail complexity and significant stakes, particularly when loan amounts exceed $250,000 and extend into the multi-million-dollar ranges.

Traditionally, global commercial and business banks have relied on experienced bankers to navigate these decisions. These bankers are supported by teams of analysts and portfolio managers who assist in credit assessment throughout the lending cycle. However, this assessment process is predominantly manual and heavily reliant on bankers’ subjective judgement. Consequently, the efficacy of banks in expanding their lending portfolios often hinges on the calibre and capacity of their banking personnel.

In the era of data and artificial intelligence (AI), there arises the opportunity for AI to augment bankers, credit analysts and portfolio managers in making well-informed decisions with greater efficiency by providing deeper insights into customer financial behaviours. This is the core challenge Rich Data Co (RDC) endeavours to address.

Termed ‘Teachable Intelligence’, our approach revolves around AI’s capacity to learn from human professionals – bankers, credit analysts and portfolio managers – and to empower them in making enhanced decisions. This symbiotic interaction between AI and human expertise not only demystifies the decision-making process but also fosters trust. We envisage a synergy of knowledge management, machine learning and generative AI working in unison to achieve this goal.

Implementing AI within Commercial and Business banks has encountered various fundamental challenges:

  • Limited Data Availability: Unlike consumer lending, where data volumes can reach millions of records, commercial lending typically involves hundreds of records, further fragmented by industry and region segmentation. Such segmentation is essential due to the distinct financial behaviours exhibited by different industries. For instance, the financial practices of agriculture are vastly different from those of retail enterprises.
  • Data Complexity: A comprehensive grasp of a business’s financial behaviour necessitates access to a broad spectrum of data. However, this data is often dispersed across multiple systems within a bank 
  • Stringent Regulatory Requirements: The lending sector, particularly in banking, operates within a highly regulated environment. Any AI model deployed in the lending domain must adhere rigorously to the established model governance framework.
  • Knowledge and Experience Capture Challenges: Bankers’ expertise is pivotal in lending decisions. How can AI effectively assimilate knowledge from bankers and provide them with meaningful support?

RDC has developed its AI decisioning platform to facilitate bankers in efficiently and securely making high-quality, intricate decisions. Some features in the platform include:

  • Robust Data Infrastructure: The RDC platform incorporates a comprehensive data sensor, data pipeline and feature library designed to safely and efficiently ingest transaction data. This allows for nuanced understanding of customer financial behaviour. By leveraging data from customers’ transaction accounts, lending accounts and financial statements, the platform enables highly predictive AI capabilities in lending.
  • Integration of Machine Learning and Knowledge Management: Recognising the limitations of machine learning when dealing with small datasets, RDC’s research has illustrated that combining machine learning with knowledge management yields more accurate predictions in lending scenarios.
  • Combining Rule-Based Decision Systems with AI: Lending decisions are often governed by numerous banking policy rules. For instance, banks may have policies prohibiting lending to businesses with a history of bankruptcy. RDC seamlessly integrates rule-based systems with AI, simplifying the design and implementation of lending decisions to enhance efficiency.
  • Transparent Decision-Making: Adhering to the risk governance framework within banks, each lending decision must be explainable and compliant. Within the RDC platform, decision logic is meticulously tracked and stored for ongoing monitoring and future reference, ensuring transparency and compliance.

Over the past 8 years, RDC has dedicated substantial effort to developing the foundational AI capabilities embedded within our platform, tailored specifically for the business and commercial lending sector. We are invigorated by the recent surge of interest and progress in the field of AI. Moreover, we are particularly enthusiastic about the prospect of empowering banks worldwide to extend lending services to millions of businesses. Our excitement stems from the opportunity to leverage AI to gain deeper insights into the financial behaviour of businesses, thereby facilitating more informed lending decisions on a global scale.


Stay updated via LinkedIn.

Contribute to the discussion on the future of AI and lending