Borrowing, getting credit, is an area that touches everyone in life. Once you reach an age when you can start to buy “stuff” you want now but don’t have ready capital, you invariably need to borrow.
Borrowing requires lending. Lending requires knowledge of the individual, business and has for many years taken the form of applying pre-defined criteria on which to assess the creditworthiness of the borrower.
At the core of the decision whether to lend is data. The more data you have about an individual borrower the better position you are to assess their creditworthiness. Data that outlines spending, collateral etc. is key to determining whether The Lender decides to accept an application. The manner in which this data is mined and reviewed is still at present highly manual.
If we accept that data is at the core of decision making (and it’s hard not to….) you then surely have to start thinking about how you can streamline this process to reduce the manual intervention and increase automation.
As a Lender, access to this data is available via the main credit agencies and (if a customer) accounts. The more complete, accurate and timely data you have will result in accepting or declining an application with less manual intervention. Incomplete or inaccurate data could result in “good” risk being declined….and a bad risk being accepted. Losing good risk and taking on bad? That’s not good business.
Utilising machine learning, AI can be introduced to your decision model to make swift and cognisant decisions. This can then be enhanced to include present and future economic factors such as inflation, micro/macro economic influences and even the strength of liquidity in the balance sheet.
It gets headier if you start to consider the total online social media presence of the borrower (as some companies already are). That is where we will end up. Before then, where to start?
The starting point is the automation of the process. Machine learning can be used to clean metadata, consider possible mistakes in an application and weed out falsifications.
Streamlining the process at the same time as making it transparent to borrowers is good business. Getting a credit application down to seconds will bring in business as well as reducing risk. Peter Maynard, Senior Vice President of Global Analytics at Equifax, in interview this year claimed their new “neural network improved the predictive ability of the model by up to 15 percent.” Using it to look back at recent decisions, they found that loans which were turned down could have been made safely.
So, if the administrative overhead can be reduced and the potential risk more cleanly defined then it can only be positive both the Lender and Borrower.