Can we truly get to paperless credit decisioning?
You are pre-approved in 5 minutes says the advert by a lender. Now what does that mean. Nothing much it seems. All it really says is you meet some minimum criteria, but we cannot take you at your word, so we need to see documentary evidence and pull a credit score. Two days or more and a day or two to receive the money in your account. This does not sound like progress. Back to the future!
So why is this so difficult to do? Is there a better way?
The answer lies in the process and the basis of decisions which is vastly unchanged over the last several decades. An Extra Long Application Form. Validate Identity. Verify Income. Check Credit Report. Underwrite. Fund. Monitor. Use Algorithms to model portfolio. Apply. Repeat. And expect better results each time!
You open an account you fill a form, you place a deposit you fill a form, you take a loan you fill a form. Lots of documentation, much of which is a repeat. If you are an existing customer a profile already exists. So why not limit to a few questions. Identity validation is fully automated but validating that the person providing the identity and the identity are the same requires a face to face. Or maybe not? How about using the digital Q&A process that are already in use for things like resetting passwords?
That leaves income and underwriting. Truthfully, its not income as much as payment behavior that matters. Credit checks validate payment behaviour and are automatic. Underwriting therefore is the bottleneck. This is where Machine Learning can help.
Using Machine Learning, banks and lenders can move beyond pre-approval to real approval. There are many institutions that are piloting this or are rolling this out.
Can we say Real Approved?