Innovation Center Denmark Silicon Valley wants to help transforming knowledge about artificial intelligence into impact in Denmark, benefitting Danish companies, the public sector and universities. We want to transform the intangible AI into Applied AI; making AI tangible, concrete and useful.
As a part of the 2017 Applied Artificial Intelligence Innovation Project focus has been on scouting for startups, corporations and business cases where artificial intelligence is the main driver of business value. This has led us to Affirm which is revolutionizing the loan industry through the use of AI. Founded in 2013 by Max Levchin, former co-founder of Paypal, Affirm is with $525 million USD in current funding bent on reinventing the industry for personal finance by providing loans to non-prime borrowers using AI to perform better risk assessment than traditional credit score can.
Affirm company description
“Our mission is simple: deliver honest financial products to improve lives.
At Affirm, we believe the financial industry desperately needs reinvention. Not only is the core infrastructure built with technology from the 1970s, but a dwindling number of people can say "I trust my bank to look out for me." It doesn’t have to be this way.
Affirm’s mission is to fix this problem. We are using modern technology to re-imagine and re-build core components of financial infrastructure from the ground up. We’re focusing on improving the lives of everyday consumers with less expensive, more transparent financial products.”
Picking Thoughts with Affirm
Innovation Center Denmark recently had the chance pick the thoughts of Chief of Staff & Director of International Markets at Affirm, Ryan Metcalf on AI, Affirm and their role in the upcoming AI revolution.
Your company’s mission statement is to “deliver honest financial products to improve lives”. How has this defined your approach to business and what role does AI play in helping consumers with less-than-perfect credit scores?
Well, our interests are completely aligned with the best interests of our customers; we do not use compounding interest, we do not have fees of any kind, we do not have revolving credit instruments. The only revenue we make from our customers is from the simple, precomputed interest that we disclose upfront to the customer. Therefore, we are only incentivized to lend to people that can afford to repay us in full and on time. This means that our underwriting models, in addition to outperforming the FICO score, must be as good as the loans we issue. This approach makes for a radically different departure from the finance industry at large where the old business models relied on cost shifting and backend pricing schemes to pay for poor underwriting and attempt to increase profits at the expense of customers.
Machine learning helps us do three things:
- Extend credit to more creditworthy borrowers whom are excluded from the traditional credit scoring systems like FICO at no fault of their own (1 in 10 Americans have unscorable credit reports).
- Assess ability and intent to repay more precisely, which decreases default and fraud rates.
- More accurately and fairly price risk for customers, and do it instantly at the point of sale.
In popular culture, AI is often portrayed as dangerous and many modern-day consumers are wary of giving out personal information. With the amount of information that your system has access to, have you experienced any apprehension from customers? If so, how have you addressed these concerns?
Max have addressed these questions before, and done so very well, so I’d rather refer to the these two longer interviews that he did on the Quantified Self and “Value for Privacy” and Affirm (In Data We Trust: Max Levchin Blows Up Consumer Finance) with Mixpanel.com back in May 2016.
Editor’s note: While both articles are worth the read - and we highly recommend reading them - we have taken the liberty to highlight a couple of quotes of Levchin’s from the articles;
“Everything that can be collected will eventually be collected. Two things, then, are essential. One, data that is collected needs to be secured, and there need to be some assurances around how secure it is. And two – by far the most important thing – is there needs to be transparency in what is being collected [and] from whom, so that a person can find out what is known about them, what has been recorded about them, even protectively.” (Max Levchin on the Quantified Self and “Value for Privacy”)
“There’s plenty of precedence for human relationships where trust is essential. It’s more about understanding what you’ve revealed than what it can be used for, how it can be used rather than the actual act of revealing.” (Max Levchin on the Quantified Self and “Value for Privacy”)
“What people really want is to understand that they are being honestly dealt with and that they know what the true cost [of the loan] will be. Telling people that a set of metrics – price, rate, whatever – is never going to change is very, very powerful. As soon as you establish that this data is the baseline, it has an unbelievably comforting effect on people trying to understand data.” (In Data We Trust: Max Levchin Blows Up Consumer Finance)
Affirm is aiming to overhaul the existing credit score system and open up the previously untouched, non-prime consumer section of the market using AI. What does this mean for the future of the traditional credit score system?
We are well on our way to overhauling the traditional credit score system. In fact, Affirm has its own internal scoring system that is proven to outperform the FICO score. Affirm's smarter underwriting approves up to 126% higher than industry averages. This means more consumers will gain leverage in the market, receive fairer pricing and be able to raise their standards of living by simply having access to capital they did not have before. It also means that companies (banks, retailers, etc.) that fail to keep up with the fintech revolution will be at a competitive disadvantage.
Compared to traditional lenders, how does your approach to data and AI differentiate you in the marketplace?
For one, Affirm's smarter underwriting approves up to 126% higher than industry averages. Furthermore, the streamlined checkout with Affirm increases conversion up to 20-25% for both mobile and desktop. Merchant partners see a 75% average order value (AOV) lift and we see 25 % of our users repeat-buy.
How have you organized your work with data on a company level vs the other aspects of your company?
We are a data and technology company first. 70 % of our company comprises of technical roles and even our finance team knows how to code.
We thank Ryan Metcalf from Affirm for answering some of our questions on the goals of such an exciting company as Affirm. Our mission here at Innovation Center Denmark is to engage companies with the future of technology from Silicon Valley. If you, or your company, are interested in exploring how you may utilize applied artificial intelligence in your business, do not hesitate to contact us. And remember; the future is already here, it is just not evenly distributed.
Check out Affirm on their own website.
Contact Senior Adviser Anders Christjansen for inquiries about our work here at Innovation Center Denmark and applied artificial intelligence in the financial sector.