Lenders Bet On Artificial Intelligence for Credit Scoring

Lenders Bet On Artificial Intelligence for Credit Scoring

Banks have to struggle with a lot of challenges – from issuing credit to operational risks, and technological troubles to good old fashion fraud. In addition to the risks of yesteryear, modern banks face falling long-term rates, growing fintech competition, and low profitability. In this challenging environment, savvy modern banks focus more of their attention […]

Lenders Bet On Artificial Intelligence for Credit Scoring

Banks have to struggle with a lot of challenges – from issuing credit to operational risks, and technological troubles to good old fashion fraud. In addition to the risks of yesteryear, modern banks face falling long-term rates, growing fintech competition, and low profitability. In this challenging environment, savvy modern banks focus more of their attention to mitigating risks.

Chief among these challenges are low-performing loan portfolios, which are a constant thorn in the side of lenders. For example, European non-performing loans stand above €1 trillion with more than one third of banks having NPL ratios above 10% (ECB, 2017).

This minefield of factors has driven lenders to seek out new ways to increase profits and cut funding costs in order to stay competitive.

Artificial Intelligence in Fintech: Will it take over?

AI is a powerful tool for banks, thanks to its ability to harness vast quantities of data to learn more about customer patterns and behaviors”, says Steve Ellis, head of the innovation group at Wells Fargo.

As powerful as artificial intelligence (AI) is, traditional banking is still heavily reliant on statistical methods that were developed over half a century ago. Lenders determine creditworthiness based on 20+ data points, which leave otherwise worthy customers behind.

Modern machine learning (ML) makes it possible to go much deeper when analyzing data, and allows lenders to extract valuable insights from available data patterns.

According to a McKinsey report, a number of European banks have already replaced the antiquated statistical-modeling approach with machine-learning techniques. The results speak for themselves: a 10% increase in the sale of new products, 20% savings in capital expenditures, and a 20% decline in churn.

Source: thefinancialbrand.com

The data doesn’t lie: Lenders are betting on AI. Evidence of this modern trend can be seen in numerous ‘banks and fintech collaborations’ and AI-based software releases:

  • JPMorgan Chase pioneered a Contract Intelligence platform designed to “analyze legal documents and extract important data points.”
  • American MobileBank deploys AI software to lend to thin-file millennials.
  • Canadian TD Bank uses Layer 6’s AI engine for scoring and cybersecurity.
  • Deutsche Bank came out with new AI-based equities to predict their pricing and volume more accurately.
  • Wells Fargo employs its own AI team to provide more personalized services and strengthen digital offerings.
  • Bank of America Merrill Lynch implements HighRadius’ AI solution to speed up receivables reconciliation for their large business clients.

Logistic regression is no longer the de facto standard

Nine times out of 10, logistic regression is used to build scoring models and solve classification issues. Before it can take over and provide predictive results, there’s an important step of preliminary analysis and data quality control that must be taken. If the dataset contains:

  • imperfect and missing values, outliers and unstructured data;
  • numerical and categorical values (age, income vs marital status, education);
  • raw data that doesn’t fit strict parameters(data with fractions or decimals, etc.)

data analysts will spend days (if not weeks) just to preprocess the data before it can be assessed. Cutting corners and ignoring such data may lead to the loss of valuable insight and incorrect predictions.

How modern AI/ML methods build better risk models

Today, lenders have the ability to collect more data than ever about their clients. In addition to traditional socio-demographic data, this may include transactional data, records from credit bureaus, social media, Google Analytics, as well as other non-traditional sources.

Processing and interpreting this data so that it can be used to issue loans to worthy credit seekers is where modern ML/AI methods give banks the edge they need.

Machine learning techniques like gradient boosting, random forest, or neural networks can better find hidden dependencies in a dataset, which helps to gain more accurate predictions. This assists banks in determining how collected parameters in a dataset should be weighed to predict whether borrowers will consistently repay their loans on time.

This is made possible by data signals, which define significant parameters that affect the power of a scoring model. Depending on the type of business, geography, target audience, and data authenticity, significant parameters may differ. Modern ML can determine which data points contain the desired signal.

Traditional data sources like credit bureaus still remain an important part of the process and provide the data that contain the above-mentioned signal. Unfortunately, they do not cover noteworthy market segments such as millennials, self-employed entrepreneurs, small business owners, immigrants, or the unbanked.

The team at GiniMachine carried out pilot projects to build accurate scoring models with minimal data points and without access to an applicant’s credit history. Some of the most promising and predictive parameters included the applicant’s industry and occupation, the size of their company, the total years they’d been in business, the size of their family, and data from social networks like their overall activity, as well as the quantity and quality of their connections.

The team at GiniMachine has proven that it is possible to capitalize on information about borrowers that is collected from alternative sources to accurately and efficiently assess borrower’s credibility and make effective lending decisions.

Modern ML methods can build more accurate risk models because of their capacity to:

  1. use built-in ‘raw’ data pre-processing tools
  2. find hidden dependencies of arbitrary complexity
  3. harness unstructured, big data, and data from alternative sources

The financial world, and lending businesses in particular, have seen major changes throughout the last few years. Using ML and AI in concert with traditional practices is the way forward for banks that want to remain competitive in the modern world. It’s clear that making good loans to the people of the future requires a futuristic helping hand.

Author:

Dmitry Dolgorukov is a CEO and co-founder of GiniMachine & HES, a technology entrepreneur, and an investor with over 15 years of executive experience in software development and fintech. In 2018, Dmitry was ranked as one of the top 200 Fintech leaders in Europe that contribute to the industry as influencers through action.

 

AI Credit Scoring: Fad or the Future?

GiniMachine density distribution

It had to happen. Artificial intelligence is shaking up the fintech world. The opportunity to get an in-depth analysis of client’s creditworthiness, have fast credit decisions, and score thin-file people makes AI a goldmine for many lenders today. Besides, machine learning and AI-based techniques have become affordable (like never before) to a wider audience, empowering […]

GiniMachine density distribution

It had to happen. Artificial intelligence is shaking up the fintech world.

The opportunity to get an in-depth analysis of client’s creditworthiness, have fast credit decisions, and score thin-file people makes AI a goldmine for many lenders today.

Besides, machine learning and AI-based techniques have become affordable (like never before) to a wider audience, empowering small and midsize lenders with the latest technology tools.

Why Today Lenders Ignore Traditional Scoring

For years, traditional scorecards, linear models, decision trees, and so widely-used FICO scores have played a significant part in a decision-making process. Although they don’t tell an applicant’s full history and can’t cope with big data, they are still used by the 90% of the top industry players.

However, more and more lending startups come to the conclusion that ‘your grandpa’s approach to data analysis’ is not enough. Because many traditional underwriting systems are missing out on a huge number of deserving borrowers.

To get traditional credit, you often need a credit history. However, according to ID Analytics, nearly 20% – or 45 million U.S. consumers – have no credit history or lack sufficient information to generate a credit bureau score. Sounds like a catch-22.

Is AI Making Credit Scores Better?

Ultimately, AI needs to improve the ‘underperforming loans’ metrics and optimize risks vs. returns for each loan issued. If it’s not doing this, then it fails.

The Efma report found that 58% of banking providers believe AI will (eventually) have a significant impact on the fintech industry.

Today, lenders have been using machine learning algorithms to solve problems both big and small, by making manual processes more simple, accurate, faster and less expensive.

“AI allows you to better or more accurately predict the one’s probability of default,” says Peter Maynard, SVP of enterprise analytics for Equifax. “Because each attribute can have multiple weights.”

Can AI Outperform Experts?

Today, in order to build a well-performing scoring model, you need to hire an expert with a fairly extensive and specific expertise in the field of mathematical statistics. Besides, this expert needs complex, specific, and very expensive software to build a model. In the end, you’ll spend hundreds of thousands of dollars and no one guarantees you results.

Paul Meehl, a clinical psychologist and professor of psychology at the University of Minnesota, conducted an experiment: He compared the predictive power of human experts with simple algorithms. The results showed that in all 20 cases, simple algorithms outperformed experts based on data such as past test scores and records of past processing.

However, algorithms won’t replace humans in complicated fields. Maybe someday, but not now. Today, AI is aimed to help experts. Scoring solution providers strive to ensure that machine learning techniques can be easily used by any lender without a need to hire a team of data scientists and developers.

AI Scoring as a ‘Black Box’

Due to the inability to understand how decisions are made, AI has been seen as a ‘black box’. Lenders admit that the hidden part of the underwriting process (in particular, how it makes predictions) looks unreliable and insecure. Business owners want to avoid the appearance of biased credit rejections, and quite often look for explainable and accountable solutions.

However, AI-based credit scoring is of limited interpretation because of its complexity. And this complexity is quantitative, not qualitative. Let me explain:

AI scoring is a set of simple rules (Age – 25+, gender – female, income is more than 1.000 but less than 1.200, etc.). These rules can be created by the tens of thousands, and it’s impossible to track all of them by with human skills.

GiniMachine, an AI-based Credit Scoring Solution

GiniMachine shares an IBM principle: Machines should do all the hard work, freeing people to think. The core idea of the solution lies in building predictive models of high quality at a reasonable cost.

Founded in 2016, a young but fast-growing fintech startup, GiniMachine aims to fight bad loans with AI.

GiniMachine combines lenders’ data insights with the best machine learning algorithms, supplemented with the set of heuristics and methodological findings. The software allows you to build custom scoring models in a few seconds without attracting expertise in the field of mathematical statistics and machine learning.

Today, lenders’ decisions to issue a loan are influenced by the credit score as well as general data like socio-demographic information or place of work of the borrower. Today, lenders take into account data from mobile devices, social media networks, the time of filling out the questionnaire, etc.

“Collecting and processing all these parameters costs a lot of money. However, not all of this information is necessary when drawing up a successful model for assessing the creditworthiness of the borrower. One of the tasks of GiniMachine is to determine which of the parameters are really important,” says Ivan Kovalenko, Co-founder of GiniMachine. “An important feature of our platform is that it knows how to work with raw data. As a result, the model built for each client is unique.”

Bottom line:

AI promises to bring lenders worldwide to the technology breakthrough. Armed with powerful solutions and best machine learning approaches, financial organizations can more effectively respond to the increasingly demanding customer base and ultimately increase acceptance rates.

Author:

Natalie Pavlovskaya

Natalie Pavlovskaya is CMO at GiniMachine, an AI-based solution for fighting bad loans. She has been actively involved with digital marketing since 2011 and has a broad range of expertise. She is passionate about fintech, AI, and e-commerce.