Consumers are obsessed with their credit scores. Some who have achieved the holy grail of a score above 800 have even been known to boast about it on their dating profiles. For businesses, which are scored in a range from 0 to 100, anything above 80 indicates good financial health and creditworthiness. But should they obsess about them – and boast about them on their LinkedIn profile?

The ‘credit bureau blind spot’ suggests they’re not as reliable as everyone thinks. This phenomenon refers to the inability of loan providers to accurately assess creditworthiness using legacy credit decisioning models that underpin their lending processes. Encumbered by narrow credit reports that overlook the applicant’s future financial position, instead relying on payment history and outstanding debt, these providers are unable to gain a holistic view of their financial position.

A lack of competition among credit bureaus also means that lenders typically make decisions based on the same information, restricting differentiation. This can make it difficult for businesses with a poor credit score to receive funding approval or do anything to improve their score.

Against this opaque backdrop, businesses are demanding instant access to real-time data that facilitates informed decision-making. Advances in artificial intelligence (AI) are generating opportunities for lenders to develop transparent credit decisioning models that mitigate the risk of rejecting creditworthy applicants and approving those whose finances might deteriorate.

Explainable AI to the rescue

The use of AI in finance is growing rapidly, with applications in areas such as risk management, fraud detection, and algorithmic trading – prompting the adoption of explainable AI (XAI) tools: a set of processes and methods that make AI models more explainable, intuitive, and understandable to human users without sacrificing performance or prediction accuracy.

Credit decisioning is also being elevated by this advanced technology, which is having a proven impact on credit-approval times and percentages. XAI-driven decisioning streamlines lending journeys by conducting real-time analysis of customer data to expedite credit decisions for retailers, small and medium-sized enterprises (SMEs), and corporate clients. This is achieved by aggregating structured and unstructured data from traditional sources (such as bank transaction history, credit reports, and tax returns) and overlooked sources (such as location data, telecom usage data, and utility bills).

By leveraging XAI to analyse these broad and diverse data sets, businesses can qualify new customers for credit services and determine loan limits and pricing expeditiously – and the benefits are compelling:

• Increase in revenue through higher acceptance rates, lower cost of acquisition, and better customer experience.
• Reduction in credit-loss rates by more precisely determining customers’ likelihood to default.
• Efficiency gains through highly automated data extraction and case prioritisation.
• Enhanced fraud management by replacing manual processes with automated detection and prevention tools.
• Improved customer experience by addressing applicants’ fear of rejection early in the funding journey and by providing instant and transparent decision-making processes.
• Reduces compliance risk by powering faster and more secure transactions, expediting informed decisions, and automating regulatory change management.

XAI and embedded finance

XAI is underpinning one of the hottest trends in the world of finance today: embedded finance. This seamless integration of financial services into non-financial ecosystems and environments is expected to grow at breakneck speed over the next decade: valued at $54.3 billion in 2022, it is forecast to reach $248.4 billion by 2032. This growth is being driven by an expectation among businesses for embedded finance providers to leverage XAI risk models and continuously invest in new ways to learn from traditionally overlooked sources.

Credit decisioning is a fulcrum of embedded lending: a subset of this new distributed approach to providing financial services that eliminates the need to rely on high-cost third parties – typically a financial institution – within the lending process. By integrating XAI-driven decisioning tools into this digital-first lending experience, it becomes a truly frictionless value-added service.

This is empowering nonfinancial businesses that embed lending functionality into their customer journey to make informed financial decisions directly within the context of their core non-financial applications or products. For example, Liberis offers a revenue-based lending model driven by an intelligent data engine that automatically forecasts business transaction revenues and makes a personalised and preapproved offer instantaneously – with 70% of businesses receiving their funding in less than 48 hours.

Credit bureaus typically focus almost exclusively on negatives – such as missed payments and prior defaults – meaning businesses don’t get rewarded for good financial behaviour. By leveraging XAI to factor in other data sources – such as revenues – and developing a more holistic view of a business, Liberis can apply a more positive mindset to our decisioning with a bias towards approving where possible.

Research by Bain Capital estimates that by 2021 around $12 billion in B2B loan transactions were made via embedded finance, which it expects to increase to between $50 billion and $75 billion by 2026. XAI is fuelling this exponential growth by allowing lenders to augment historic credit data with forward-looking insights into an applicant’s suitability to repay a debt obligation in the future – and present their findings transparently and expeditiously.

The future of XAI in credit decisioning

XAI is playing a vital role in helping more SMEs gain access to the credit they deserve so they can live their best financial lives – and it’s only just getting started. By its very nature, XAI is constantly developing. This dynamism will continue to enrich credit decisioning processes through advancements in open bank connectivity for access to richer data, and by harnessing counterintuitive data that challenges the status quo.