Author: Subbu Venkataramanan
Introduction
In the last few years, some business trends led Credit Unions to change how they do business. These include changes in regulatory frameworks, inflation rates, economic policies, compliance guidelines, automation, and other disruptive technologies. They not only pushed Credit Unions to adapt but also left no place for the ones who denied the change. However, technological changes turned out to be positive for the industry; they brought about newer technologies such as Artificial Intelligence (AI) and Machine Learning (ML). They use tons of data that financial institutions generate, to draw intelligence and detect patterns to create effective marketing programs, earn more revenue opportunities, build deeper relationships with their customers, and provide enhanced digital customer experience.
In this point of view report, we aim to establish the role technologies such as Artificial Intelligence can play in an ecosystem of a credit union, specifically in credit decisioning. For the last few years, it was assumed by many credit unions and others that AI/ML will be responsible for the reduction of jobs; but that assumption is changing now. Credit unions are, in fact accepting how AI/ML is helping them add customized intelligence that can enhance its member relationships and give a very personal banking experience.
Credit Unions today
The 2008 downturn changed the face of the U.S. economy, but we all saw how credit unions largely weathered that storm. According to the U.S. News and World Report in 2013, as few as 2% of credit unions in the U.S. were forced to liquidate or merge from 2008 to 2012 (versus about 7% of banks), and mortgage loan delinquency was the lowest at 1.61% vs. 8.86% at banks. These numbers confirm the standing that credit unions have in today's U.S. economy.
Credit unions are nonprofit, member-owned banking institutions that work for the benefit of its members. These days a large number of individuals and small businesses are turning towards credit unions for their credit instruments as they are considered to be trustworthy lending institutions with a more humanized approach to the business. According to the National Credit Union Association, in 2018, the country saw around 5,757 credit unions operating nationwide, serving approx. 103.992 million people. And this number suggests how our readings about the credit unions are correct. Another example to support the statement could be the 2011 grassroots-led "Move Your Money" campaign - the campaign witnessed credit unions add more than $400 billion worth of new consumer deposits. Higher deposits allowed credit unions to fund more loans, reinvest profits to provide direct benefits to its members, and offer more modern savings-oriented products. Mainly, these are just some of the reasons why credit unions are a preferable option for many.
While credit unions remain to be the most trusted lending institution today, technology evolution is coming comparatively slowly to them. For the last few years, credit unions amongst many were hesitant to look at Artificial Intelligence or Machine Learning to optimize their existing ecosystem. The stigma around AI/ML had left many companies wondering whether implementing these newer technologies will lead to a robotic world with no human jobs left to do. But slowly, that mindset is changing. In the past two years, we have seen many large credit unions getting into strategic partnerships with FinTech’s that can implement these technologies for them. Implementing AI/ML is transforming how lenders of all sizes market, underwrite, and service their customers. Credit unions have the chance to embrace this game-changing opportunity to be competitive in the lending market and deliver best-in-class consumer experience to its members.
For credit unions, Artificial Intelligence is an opportunity to deepen those already strong member relationships. And hence it will be beneficial for credit unions to see how AI can help solve their business problems and enhances member experiences.
Why is Artificial Intelligence important for Credit Unions?
Artificial Intelligence uses alternative data that includes information that isn't typically found in consumers' credit reports and isn't necessarily provided by the consumer at the time of loan application. It may consist of members' behaviors, financial patterns, digital activity on products, and financial history, which could lead to producing intelligent insights that credit unions can use to deepen their relationships with its members. Artificial Intelligence identifies if a specific service/offer is relevant to a member and then triggers the delivery of that offer/service/product. Some of the suggestions could be personal credit, savings, loans, insurance, and a home mortgage.
Implementing Artificial Intelligence at credit unions can allow them to:
Offer a new service/product to a member at the right time
Provide a more personalized customer service experience especially when there's an issue
Deliver better convenience to delight your customers
We can already see the consumers demanding these experiences from their mortgage lenders and banks, and it stands true for credit unions too. To remain competitive in this changing market, credit unions must act swiftly towards providing these experiences to its members. Essentially, credit unions must let their data work for them. Deployed successfully, AI will help credit unions in deepening the already strong relationships among members.
Where can Credit Unions focus their AI Implementation?
Currently, AI implementations in credit unions include:
Member Onboarding
Loan Application and Risk Decisions in credit Underwriting
Customer Service
Credit Scoring
Digital Experience including Mobile and Web
New Offer/Service/Product Recommendations
AI offers unique insights into each of these areas that credit unions can benefit from in their business. Now, let's look at some of the primary focus areas that credit unions are considering implementing AI on:
Better decisions in credit underwriting: Some FinTechs offer AI-powered underwriting models, that can utilize data, make elaborate mathematical calculations, and automation to evaluate and prompt the risk that a new or an existing member seeking a loan carries. With AI-powered underwriting, credit unions can become able to assess an applicant almost immediately. This will allow credit unions to deliver a superior experience to its members while closing more loans. This also reduces the duration of discussions and increases cross-sell opportunities for other lending products. AI will also help credit unions flag the high-risk applications that an underwriter can verify in detail and can even engage those members in a conversation. AI-powered underwriting will help loan officers allocate their time to engaging with a client more intelligently than before.
Improved understanding of members: Members are the pillars of strength for credit unions, and as long as they are satisfied and safe, credit unions' work stays robust. AI combined with Machine Learning can be implemented to analyze the vast data available with credit unions to gain essential insights on members. For example, based on the patterns in cash-flows and buying behavior, if a credit union can identify the birth of a child in a member's family, then communication can be triggered to propose a plan for safeguarding a child's' future. Additionally, many advanced AI-powered security solutions can be deployed to make sure that members' data is safe.
Increased lending opportunities: Traditionally, banks have had the required resources (real estate, budgets, etc.) at their disposal to invest in the necessary technology to compete effectively in the market. But recently, we have witnessed a shift in the trend - for instance, a FinTech named "Mirador" came up with a lending platform that uses Machine Learning and Predictive Analytics to help credit unions compete against larger financial institutions in originating to small businesses. It has put the focus back on smaller loans that may not be very attractive to the big banks. Lending smaller amounts may turn into a big profit-making business model for credit unions. Growth in lending opportunities is possible with the use of Artificial Intelligence and make the loan cycle more accessible and personalized.
Enhanced customer service: The world already knows AI in the form of Siri, Alexa, and Okay Google! - and web/mobile chat-bots are not new either. But, chat-bots are now AI-powered, customized to one's digital activity, and answers in a more humane-way to give you a more personalized experience. A good chat-bot can authenticate members and potentially help them with their accounts.
Additionally, credit unions can utilize and deploy AI-enabled, 24/7 customer interactions, and service. It can answer hundreds of queries from hundreds of members simultaneously, providing instant access to vital information to avoid repetition and a chance for credit unions to connect with their members well.
CFPB and other federal agencies encourage the use of alternative data
As published on CFPB Newsroom, five federal financial regulatory agencies issued a joint statement that the use of alternative data in underwriting by banks, credit unions, and non-bank financial firms to evaluate creditworthiness may be a way to increase access to credit or decrease the cost of the loan.
The statement from the Federal Reserve Board (Federal Reserve), the Consumer Financial Protection Bureau (CFPB), the Federal Deposit Insurance Corporation (FDIC), the Office of the Comptroller of the Currency (OCC), and the National Credit Union Administration (NCUA) highlights the benefits that using alternative data may provide to consumers, such as expanding access to credit and enabling consumers to obtain additional products and more favorable pricing and terms. The statement explains that a well-designed compliance management program provides for a thorough analysis of relevant consumer protection laws and regulations to ensure that firms understand the opportunities, risks, and compliance requirements before using alternative data.
2019 CFPB statement and their study clearly show that alternative data can lead to 27% more loan approvals. We see how a large number of lenders or other lending companies still rely on age-old blunt credit instruments while incurring $130 Bn credit losses annually. The time is ripe to change that by implementing AI and ML to gain more lending opportunities as opposed to bearing losses.
There are many AI solutions available in the marketplace, which can be implemented via API integration or with a systematic partnership or can be acquired. However, the implementation process includes strategy building and model deployment, which is complicated, expensive, and takes months. Implementing AI-solutions is not an easy task for a credit union. It needs to be left to experts as it requires a solution that puts its components and processes into perspective and makes it work together for the benefit of credit unions and its members.
AI-based credit underwriting: The way forward
Sharper AI, ML credit underwriting tools offer better credit decisions. This takes care of strategy, AI-ML driven modeling, and enables real-time decisions and also renders the deployment process a lot simpler and faster.
1: Data
With sharper tools, the power of internal data can be unlocked even before the implementation of alternative data into the decisioning process. It involves bringing together silos of data and configure appropriate features. However, there is a need to bring alternative data sources and essential features to augment internal data for underwriting. The richer the data and features/signals derived from the data, the richer the quality of underwriting.
2. AI/ML driven modelling
AI-ML algorithms can unearth unapparent patterns which were not considered in the decisioning process. The ability of AI-ML tools enables decisioning to be more customer centric and close to real time. Simplified creation and management of explainable AI models will improve the quality of underwriting decisions and improve loan approval rates while bringing down defaults.
How to implement AI/ML models:
3: Enable real-time decisioning using AI/ML
Having said that, just data and algorithms will not be of much use if they cannot be implemented quickly and effectively. The deployment of technology should try to fit into existing systems and processes to aid quick integration.
There's a specific need for:
Full lifecycle solution which integrates into real-time and batch process
Prebuilt predictors & models in a platform for rapid deployment & ROI
Flexibly strategy builder
Flexible architecture that integrates with existing processes
Such an environment will accelerate the deployment of new-age credit decisions without disrupting existing processes and technologies.