There’s been a lot of discussion on the conference circuit about how artificial intelligence (AI) and machine learning (ML) can drive financial services forward as we enter a new decade. It’s time for banks and financial institutions to optimise the signal-to-noise (another AI technology) and find practical applications for the latest AI / ML technology, which can unlock a range of benefits for lenders.
Whilst alternative lenders like Lending Club use AI / ML on a daily basis to inform loan decisions, traditional financial institutions have stayed away and watched curiously from the side-lines, driven by supranational regulatory requirements that the executives of a bank or other lender understand the underlying factors at play in any automated lending decision.
Banks have continued to rely on analogue factors like credit policies and debt-to-income ratios for underwriting, even though they know they must embrace the digital age and that, in many applications of the latter, this has no impact on prediction of default.
They must take a leap of faith. A new study by IHS Markit suggests the business value of AI / ML to the banking sector will reach $300 billion by 2030, whilst Gartner believes that income from AI / ML will triple over the next few years.
The big advantages of using AI / ML systems are clear:
- Routine processes become automated
- The speed of service increases
- Costs for decisioning decrease
- Accuracy of data processing increases
And you get better credit insights
Jackson Mueller from the Center for Financial Markets at the Milken Institute has pointed to multiple studies that highlight the benefits of AI / ML when it comes to lending, including a detailed report that found some of the algorithms used by fintech lenders discriminate 40% less than face-to-face lenders when it comes to assessing creditworthiness by applying the objectivity associated with AI / ML.
Many of the complex algorithms used in machine learning can assess non-numerical factors in creditworthiness evaluation, including consumer behaviour in other industries. Cutting-edge credit scoring capabilities powered by AI / ML can offer greater predictive power when leveraged on an applicant's willingness and ability to repay debt and help to extend credit to applicants who might otherwise have been excluded.
Appropriate use of AI / ML in loan origination also reduces the human error in terms of processing applications or missing critical factors in whether a borrower will default on a loan. It is anticipated that AI / ML will play an increasing role in banks’ loan management systems in helping to identify patterns of behaviour that indicate a consumer may be close to declaring insolvency or default, thus reducing expected credit losses.
Lessons from big tech
Lenders should consider how their competitors harness AI / ML. Amazon, for example, holds proprietary information on what products are sold on its site, how customers rate those products, the economic status of those companies which manufacture and distribute those products, and predicting future demand for these products.
Amazon is using this wealth of data to develop AI / ML models in order to find companies to offer small business loans to. It selects the companies and makes the application journey easy. Last year, Amazon loaned around $1 billion to small businesses that use its marketplace.
Get in touch today to see how ezbob can help you make the shift from analogue to digital.