How a Norwegian virtual bank is using machine learning

The robot uprising is coming, even to your loan application process. Norway's Instabank is appealing to customers who want to get their banking done outside a brand setting by using a machine learning and artificial intelligence-based platform. "We can accept applications from a user, and give the [loan] offer and pay out on the same day," said Instabank chief marketing officer Lauren Pedersen. "Everything can be done digitally, while a lot of the incumbents require paperwork -- it comes back to speed, efficiency and convenience." Through the help of cloud-based, risk-decisioning platform developed by technology company Provenir, Instabank, a digital bank with a 20-person staff, is able to evaluate customers in near real time and have funds transferred to a customer's bank account within a day. The tool allows Instabank to gather the data through an online questionnaire that asks traditional loan eligibility questions, including a credit check, current debt and income. Pedersen said what sets it apart from some larger players is that's processed quickly to result in a near-immediate decision, and the bank is able to make changes to the risk assessment in real time to keep pace with changing customer behavior and needs. "We want to be able to do optimizations, updates and changes really fast; when we see that there's a change that we want to make from our analysis [of a customer's risk profile], we want to be able to make that change and launch the same day, and that's something other banks will take a long time on their side to be able to update and change." Instabank has picked up customers fast; since it began operations last year, it's gained over 7,000 customers. But when it comes to loans, banks in Norway are moving away from lengthy paperwork. For example, DNB, Nordea and SpareBank 1 offer online loan applications. This is perhaps due in part to the country's transition to a cashless culture, as shown by the government's moves to eliminate cash by 2030 and require mandatory online billing. But according to the technology company that developed the platform that Instabank uses, its ability to use artificial intelligence and machine learning to easily crawl through structured and unstructured data sources is a distinguishing factor. "We can pull data from identifying information based on any rules [the bank customer] has set, and we can determine whether or not they're creditworthy," said JoAnn Martin, head of content marketing and automation for Provenir, which works with financial services companies around the world, including Nordea, Elavon, Klarna, HSBC and Wells Fargo. For instance, instead of manually determining an age or address cutoff to determine loan eligibility, the machine-learning platform could be trained to sift through historical salary or payment data to offer the best risk outcome. While Instabank said it sticks to traditional criteria for assessing creditworthiness, Martin said Provenir's capabilities can be used to stretch traditional underwriting methods to incorporate social media data. It's a feature some of its customers use, but Martin wouldn't say which ones. "Imagine a small business is applying for a loan and they want some sort of pre-qualification from an online lender; the lender can look at a Yelp profile and say that a profile that's been around three months and has 100 terrible reviews and zero good ones can be a determining factor," she said, noting "it's up to our customers to set the rules -- banks have their own models they use."