What Banks Can Learn About Customers from 50,000 Chatbot Searches

Covid-19 has increased usage of digital banking services and tools, including live chat, video chat and chatbots.

While live chat and video chat offer a one-to-one conversation directly with your customers, chatbots provide 24-hour service, instant answers and the ability to scale without the need for human intervention. Relatively new channel to the banking world, the promise of chatbots seems endless: answering every question and automating related tasks, quickly and efficiently. How can banks best leverage the promise of this opportunities to better and more efficiently serve customers?

To truly answer that question, we need to understand how customers interact with chatbots, how that varies from known digital behavior, like search and navigation, and how can those insights be turned into reality.

So we decided to analyze more than 50,000 banking chatbot interactions. What we uncovered revealed some very interesting insights about customer behavior and what it will take to make that promise a reality.

It turns out that customers interactions with chatbots are very similar to human interactions:

  • They typically typed 11.24 words, on average, compared to with 1.4 words typed into a banking website search bar. Chatbot interactions are conversational. Customers ask questions like “Can I Have My Stimulus Debit Card Balance Deposited to My Account” or making statements like “I need to change my address.”
  • Almost 94% of questions asked were completely unique. While customers may ask the same type of question — “What is my routing number?” versus “What is your routing number?” versus “What is the routing number” —how they phrase the question is almost always unique.
  • A fifth of all interactions started with “I need,” “I want” or “I am” — another indication of the conversational approach that bank customers take with chatbots. Unlike a search function, where typically they would use shorter phrases like “refinance” or “refinance car,” they make statements or ask questions: “I am looking to refinance my auto loan” or “I want to refinance my auto loan.”
  • Fifteen percent of interactions included the word “how.” This is another indication that customers ask chatbots questions or looking for help completing tasks like “How do you use Zelle?” or “How does a home equity loan work?”
  • Fourteen percent of all interactions began with “Hi,” “Hey” or “Hello.” And who said that bots don’t have feelings?

Chatbot adoption and usage will only continue to grow. Like all newer channels, it will require fine-tuning along the way, using insights and analysis to effectively interpret what customers are looking for, and deliver back relevant responses that point them in the right direction.

This starts with analytics and data. As data sets grow with more usage, they will reveal insights on how customers interact with chatbots, what they are looking to do and how that changes over time. This will feed the data set used to power the chatbot’s AI — both natural language processing (the ability to interpret what the customer is asking or looking for) as well as the sentiment analysis (whether the customer is happy or frustrated). Analysis will be required to learn and understand the nuances of what customers are asking when presented with phrases like this actual query from our dataset: “Hi. What is the safest way to prove documents of account balance when applying to living in an apartment complex?” Banks and/or the chatbot vendors will need to monitor the training the chatbot, including recognizing customer frustration and offering up logical next steps — like “It looks like you’re frustrated, can we transfer you to an agent?” as needed.

The analytics and data will also provide the map of the information that needs to be developed and updated to deliver answers that customers need. Given that 93.8% of questions that customers ask are unique, having the right knowledge will be critical. Sometimes this might be a simple answer (“What is my routing number?”) and sometimes it might require decision trees that offer options (understanding if an auto loan is for a new or used vehicle to get the customer one step closer to conversion).

Banks have a great opportunity to make chatbots the 24/7 tool that improves customer experience, reduces support costs and drives digital adoption. But it will take a commitment to the analysis and ongoing optimization of knowledge to truly become a reality. 

Next time you start you interact with a chatbot, start with hello — I’ve heard they appreciate it!

Five Ways PPP Accelerates Commercial Lending Digitization

The Small Business Administration’s Paycheck Protection Program challenged over 5,000 U.S. banks to serve commercial loan clients remotely with extremely quick turnaround time: three to 10 days from application to funding. Many banks turned to the internet to accept and process the tsunami of applications received, with a number of banks standing up online loan applications in just several days. In fact, PPP banks processed 25 times more loan applications in 10 days than the SBA had processed in all of 2019. In this first phase of PPP, spanning April 3 to 16, banks approved 1.6 million applications and distributed $342 billion of loan proceeds.

At banks that stood up an online platform quickly, client needs drove innovation. As institutions continue down this innovation track, there are five key technology areas demonstrated by PPP that can provide immediate value to a commercial lending business.

Document Management: Speed, Security, Decreased Risk
PPP online applications typically provided a secure document upload feature for clients to submit the required payroll documentation. This feature provided speed and security to clients, as well as organization for lenders. Digitized documents in a centrally located repository allowed appropriate bank staff easy access with automatic archival. Ultimately, such an online document management “vault” populated by the client will continue to improve bank efficiency while decreasing risk.

Electronic Signatures: Speed, Organization, Audit Trail
Without the ability to do in-person closings or wait for “wet signature” documents to be delivered, PPP applications leveraged electronic signature services like DocuSign or AdobeSign. These services provided speed and security as well as a detailed audit trail. Fairly inexpensive relative to the value provided, the electronic signature movement has hit all industries working remotely during COVID-19 and is clearly here to stay.

Covenant Tickler Management: Organization, Efficiency, Compliance
Tracking covenants for commercial loans has always been a balance between managing an existing book of business while also generating loan growth. Once banks digitize borrower information, however, it becomes much easier to create ticklers and automate tracking management. Automation can allow banker administrative time to be turned toward more client-focused activities, especially when integrated with a document management system and electronic signatures. While many banks have already pursued covenant tickler systems, PPP’s forgiveness period is pushing banks into more technology-enabled loan monitoring overall.

Straight-Through Processing: Efficiency, Accuracy, Cost Saves
Banks can gain significant efficiencies from straight-through processing, when data is captured digitally at application. Full straight-through processing is certainly not a standard in commercial lending; however, PPP showed lenders that small components of automation can provide major efficiency gains. Banks that built APIs or used “bots” to connect to SBA’s eTran system for PPP loan approval processed at a much greater volume overall. In traditional commercial lending, it is possible for data elements to flow from an online application through underwriting to final entry in the core system. Such straight-through processing is becoming easier through open banking, spelling the future in terms of efficiency and cost savings.

Process Optimization: Efficiency, Cost Saves
PPP banks monitored applications and approvals on a daily and weekly basis. Having applications in a dynamic online system allowed for good internal and external reporting on the success of the high-profile program. However, such monitoring also highlighted problems and bottlenecks in a bank’s approval process — bandwidth, staffing, external vendors and even SBA systems were all potential limiters. Technology-enabled application and underwriting allows all elements of the loan approval process to be analyzed for efficiency. Going forward, a digitized process should allow a bank to examine its operations for the most client-friendly experience that is also the most cost and risk efficient.

Finally, these five technology value propositions highlight that the client experience is paramount. PPP online applications were driven by the necessity for the client to have remote and speedy access to emergency funding. That theme should carry through to commercial banking in the next decade. Anything that drives a better client experience while still providing a safe and sound operating bank should win the day. These five key value propositions do exactly that — and should continue to drive banking in the future.

Build Versus Buy Considerations for Data Analytics Projects

It is the age-old question: buy versus build? How do you know which is the best approach for your institution?

For years, bankers have known their data is a significant untapped asset, but lacked the resources or guidance to solve their data challenges. The coronavirus crisis has made it increasingly apparent that outdated methods of distributing reports and information do not work well in a remote work environment.

As a former banker who has made the recent transition to a “software as a service” company, my answer today differs greatly from the one I would have provided five years ago. I’ve grown in my understanding of the benefits, challenges, roadblocks and costs associated with building a data analytics solution.

How will you solve the data conundrum? Some bank leaders are looking to their IT department while other executives are seeking fintech for a solution. If data analytics is on your strategic roadmap, here are some insights that could aid in your decision-making. A good place to start this decision journey is with a business case analysis that considers:

  • What does the bank want to achieve or solve?
  • Who are the users of the information?
  • Who is currently creating reports, charts and graphs in the institution today? Is this a siloed activity?
  • What is the timeline for the project?
  • How much will this initiative cost?
  • How unique are the bank’s needs and issues to solve?

Assessing how much time is spent creating meaningful reports and whether that is the best use of a specific employee’s time is critical to the evaluation. In many cases, highly compensated individuals spend hours creating reports and dashboards, leaving them with little time for analyzing the information and acting on the conclusions from the reports. In institutions where this reporting is done in silos across multiple departments and business units, a single source of truth is often a primary motivator for expanding data capabilities.

Prebuilt tools typically offer banks a faster deployment time, yielding a quicker readiness for use in the bank’s data strategy, along with a lower upfront cost compared to hiring developers. Vendors often employ specialized technical resources, minimizing ongoing system administration and eliminating internal turnover risk that can plague “in house” development. Many of these providers use secure cloud technology that is faster and cheaper, and takes responsibility for integration issues.  

Purchased software is updated regularly with ongoing maintenance, functionality and new features to remain competitive, using feedback and experiences gained from working with institutions of varying size and complexity. Engaging a vendor can also free up the internal team’s resources so they can focus on the data use strategy and analyzing data following implementation. Purchased solutions typically promotes accessibility throughout the institution, allowing for broad usage.

But selecting the criteria is a critical and potentially time-consuming endeavor. Vendors may also offer limited customization options and pose potential for integration issues. Additionally, time-based subscriptions and licenses may experience cost growth over time; pricing based on users could make adoption across the institution more costly, lessening the overall effectiveness.

Building a data analytics tool offers the ability to customize and prioritize development efforts based on a bank’s specific needs; controllable data security, depending on what tools the bank uses for the build and warehousing; and a more readily modifiable budget.

But software development is not your bank’s core business. Building a solution could incur significant upfront and ongoing cost to develop; purchased tools appear to have a large price tag, but building a tool incurs often-overlooked costs like the cost of internal subject matter experts to guide development efforts, ongoing maintenance costs and the unknowns associated with software development. These project may require business intelligence and software development expertise, which can carry turnover risk if institutional knowledge leaves the bank.

Projects of this magnitude require continuous engagement from management subject matter experts. Bankers needed to provide the vision and banking content for the product — diverting management’s focus from other responsibilities. This can have a negative impact on company productivity.

Additionally, “in-house” created tools tend to continue to operate in data silos whereby the tool is accessible only to data team. Ongoing development and releases may be difficult for an internal team to manage, given their limited time and resources along with changing business priorities and staff turnover.

The question remains: Do you have the bandwidth and talent at your bank to take on a build project? These projects typically take longer than expected, experience budget overruns and often do not result in the desired business result. Your bank will need to make the choice that is best for your institution.

Five Digital Banking Initiatives for Second Half of 2020

As the calendar nears the midpoint of 2020 and banks continue adjusting to a new normal, it’s more important than ever to keep pace with planned initiatives.

To get a better understanding of what financial institutions are focusing on, MX surveyed more than 400 financial institution clients for their top initiatives this year and beyond. We believe these priorities will gain even more importance across the industry.

1. Enabling Emerging Technologies, Continued Innovation
Nearly 20% of clients see digital and mobile as their top initiatives for the coming years. Digital and mobile initiatives can help banks limit the traffic into physical locations, as well as reduce volume to your call centers. Your employees can focus on more complex cases or on better alternatives for customers.

Data-led digital experiences allow you to promote attractive interest rates, keep customers informed about upcoming payments and empower them to budget and track expenses in simple and intuitive ways. 

2. Improving Analytics, Insights
Knowing how to leverage data to make smarter business decisions is a key focus for financial institutions; 22% of our clients say this is the top initiative for them this year. There are endless ways to leverage data to serve customers better and become a more strategic organization.

Data insights can indicate customers in industries that are at risk of job loss or layoffs or the concentration of customers who are already in financial crisis or will be if their income stops, using key income, spending and savings ratios. Foreseeing who might be at risk financially can help you be proactive in offering solutions to minimize the long-term impact for both your customers and your institution.

3. Increasing Customer Engagement
Improving and increasing customer engagement is a top priority for 14% of our clients. Financial institutions are well positioned to become advocates for their customers by helping them with the right tools and technologies.

Transaction analytics is one foundational tool for understanding customer behavior and patterns. The insights derived from transactions and customer data can show customers how they can reduce unnecessary spending through personal financial management and expert guidance.

But it’s crucial to offer a great user experience in all your customer-facing tools and technologies. Consumers have become savvier in the way they use and interact with digital channels and apps and expect that experience from your organization. Intuitive, simple, and functional applications could be the difference between your customers choosing your financial institution or switching to a different provider.

4. Leveraging Open Banking, API Partnerships
Open banking and application programming interfaces, or APIs, are fast becoming a new norm in financial services. The future of banking may very well depend on it. Our findings show that 15% of clients are considering these types of solutions as their main initiative this year. Third-party relationships can help financial institutions go to market faster with innovative technologies, can strengthen the customer experience and compete more effectively with big banks and challengers.

Financial institutions can leverage third parties for their agile approach and rapid innovation, allowing them to allocate resources more strategically, expand lines of business, and reduce errors in production. These new innovations will help your financial institution compete more effectively and gives customers better, smarter and more advanced tools to manage their financial lives.

But not all partnerships are created equally. The Office of the Comptroller of the Currency recently released changes surrounding third-party relationships, security and use of customers’ data, requiring financial institutions to provide third-party traffic reports of companies that scrape data. Right now, the vast majority of institutions only have scrape-based connections as the means for customers to give access to their data — another reason why financial institutions should be selective and strategic with third-party providers.

5. Strategically Growing Customer Acquisition, Accounts
As banking continues to transform, so will the need to adapt including the way we grow. Nearly 30% of our clients see this as a primary goal for 2020 and beyond. Growth is a foundational part of success for every organization. And financial institutions generally have relied on the same model for growth: customer acquisitions, increasing accounts and deposits and loan origination. However, the methods to accomplish these growth strategies are changing, and they’re changing fast.

Right now, we’re being faced with one of the hardest times in recent history. The pandemic has fundamentally changed how we do business, halting our day-to-day lives. As we continue to navigate this new environment, financial institutions should lean on strategic partnerships to help fill gaps to facilitate greater focus on their customers.

Helping Customers When They Need It Most

Orvin Kimbrough intimately understands the struggles shared by low-to-moderate income consumers. Raised in low-income communities and the foster care system, he also worked at the United Way of Greater St. Louis for over a decade before joining $2.1 billion Midwest BankCentre as CEO in January 2019. “[Poverty costs] more for working people,” he says. “It’s not just the financial cost; it’s the psychological cost of signing over … the one family asset you have to the pawn shop.”

His experiences led him to challenge his team to develop a payday loan alternative that wouldn’t trap people in a never-ending debt cycle. The interest rate ranges from 18.99% to 24.99%, based on the term, amount borrowed (from $100 to $1,000) and the applicant’s credit score. Rates for a payday loan, by comparison, range in the triple digits.

The application process isn’t overly high-tech, as applicants can apply online or over the phone. The St. Louis-based bank examines the customer’s credit score and income in making the loan decision; those with a credit score below 620 must enroll in a financial education class provided by the bank.

Industry research consistently finds that many Americans don’t have money saved for an emergency — a health crisis or home repair, for example. When these small personal crises occur, cash-strapped consumers have limited options. Few banks offer small-dollar loans, dissuaded by profitability and regulatory constraints following the 2008-09 financial crisis.

If the current recession deepens, more consumers could be looking for payday loan alternatives. Regulators recently encouraged financial institutions to offer these products, issuing interagency small-dollar lending principles in May that emphasize consumers’ ability to repay. 

Everybody needs to belong to a financial institution if you’re going to be financially healthy and achieve your financial aspirations,” says Ben Morales, CEO of QCash Financial, a lending platform that helps financial institutions automate the underwriting process for small-dollar loans.

QCash connects to a bank’s core systems to automate the lending process, using data-driven models to efficiently deliver small-dollar loans. The whole process takes “six clicks and 60 seconds, and nobody has to touch it,” Morales says. QCash uses the bank’s customer data to predict ability to repay and incorporates numerous factors — including cash-flow data — into the predictive models it developed with data scientists. It doesn’t pull credit reports.

Credit bureau data doesn’t provide a full picture of the customer, says Kelly Thompson Cochran, deputy director of FinRegLab and a former regulator with the Consumer Financial Protection Bureau. Roughly a fifth of U.S. consumers lack credit history data, she says, which focuses on certain types of credit and expenses. The data is also a lagging indicator since it’s focused on the customer’s financial history.

In contrast, cash flow data can provide tremendous value to the underwriting process. “A transaction account is giving you both a sense of inflows and outflows, and the full spectrum of the kind of recurring expenses that a consumer has,” says Cochran.

U.S. Bancorp blends cash flow data with the applicant’s credit score to underwrite its “Simple Loan” — the only small-dollar loan offered by a major U.S. bank. The entire process occurs through the bank’s online or mobile channels, and takes just seven minutes, according to Mike Shepard, U.S. Bank’s senior vice president, consumer lending product and risk strategy. Applicants need to have a checking account with the bank for at least three months, with recurring deposits, so the bank can establish a relationship and understand the customer’s spending behavior.

“We know that our customers, at any point in time, could be facing short-term, cash-flow liquidity challenges,” says Shepard. U.S. Bank wanted to create a product that was simple to understand, with a clear pricing structure and guidelines. Customers can borrow in $100 increments, from $100 to $1,000, and pay a $6 fee for every $100 borrowed. U.S. Bank lowered the fee in March to better assist customers impacted by the pandemic; prior to that the fee ranged from $12 to $15.

Since the loan is a digital product, it’s convenient for the customer and efficient for the bank.

Ultimately, the Simple Loan places U.S. Bank at the center of its customers’ financial lives, says Shepard. By offering a responsible, transparent solution, customers “have a greater perception of U.S. Bank as a result of the fact that we were able to help them out in that time of need.”

Using Data Platforms to See Customers

Customers leave behind valuable breadcrumbs about their interests, needs and intentions across their financial lives.

What’s their current financial health? Are they shopping for a new credit card? Even: Are they considering switching to a competitor?

Unfortunately, this wealth of insights is more-than-likely locked away across a series of legacy, on-site systems, stuck in siloed data warehouses and generally difficult to access due to antiquated reporting systems. Understanding and acting on customer signals has become more important in recent months as customers seek financial partners that understand their unique needs. What does it take for a bank to unlock this treasure trove of data and insights? More often than not, a customer data platform (or CDP) can help banks take an important step in making this a reality and craft a 360-degree view of their customers.

I spoke with Brian Knollenberg, vice president of digital marketing and analytics at Tukwila, Washington-based BECU, about his recent experience of setting up a CDP for one of the country’s largest credit unions in the country. 

The Need for CDP
When Knollenberg joined the $22 billion credit union, he saw that creating a marketing performance dashboard using slow-batch processing across multiple systems took 12 manual hours to produce. As a result, the data stakeholders needed to make key decisions was a week out of date by the time they received it — much less take action on it.

This speed-to-value lag wasn’t limited to just marketing dashboards; it was just one example teams encountered when trying to access timely customer data across legacy systems. His team recognized that the organization needed current data, individualized for each customer, to make timely decisions. They also needed a way to easily syndicate this across critical customer and stakeholder touchpoints. 

Knollenberg also recognized his team’s expertise was better suited to modifying processes rather than building a robust enterprise-grade tool that could ingest and process terabytes of data in near-real time. He needed a solution to transform this data hindrance into an asset, and looked for a partner with direct experience in tackling these challenges to streamline implementation.

CDP Benefits
Implementing a CDP has extended the BECU team’s ability to tackle more difficult data challenges. This included building out performance dashboards that update every 24 hours, personalized customer communications and the ability to modeling member financial health.

This last use case empowers BECU to aggregate a score based on behaviors, transactions, and trends to identify which members could benefit from proactive outreach or help. He said financial health scoring has been extremely helpful during the coronavirus pandemic to identify potential recipients of proactive outreach and assistance. Having this information readily available enables marketing, customer service and even product teams to create bespoke experiences for their members and make informed business decisions — like offering a lower rate card to a member showing large carried balances with an outside card provider.

Lessons Learned
Before tackling any new data program, Knollenberg recommends companies first identify the overall effort versus impact. He finds that while companies often invest ample time and effort into developing comprehensive strategy and goals, they often miss when planning for the execution realities to properly implement them. Spend time scaling up your bank’s execution capabilities, determine how you’ll realistically measure potential impact and test-drive product solutions via a robust proof of concept.

The best financial brands know that putting their customers first will result in returns. Building out a customer data platform for your bank can unlock powerful new insights and opportunities to engage with your customers, if done right. As you start on this journey, make sure to identify what specific use cases are most impactful for your business, and find the right software partner that will work with you to execute it properly. Once unlocked, your bank will be able to service customers at a truly personalized level and drive a greater share of wallet.

Three Financial Institutions, Three Ways to Improve Operations

cinchy.pngIt can be hard for banks to invest in changing their operations, especially when new systems or vendors are involved. Unlike loans or accounts, assigning a price tag to the value banks get from improving their operations is far from straightforward.

With this in mind, Bank Director looked at how three fintechs — Cinchy, Empyrean Solutions and INETCO Systems Limited — have helped three financial institutions streamline data sharing, financial modeling and real time transaction modeling.

Cinchy, a data collaboration platform that manages data as a network and enables banks to build their own business applications, won Best Solution for Improving Operations at Bank Director’s 2020 Best of FinXTech Awards in May. Modeling platform Empyrean Solutions and transactions intelligence firm INETCO Systems Limited were finalists for the category.

Cinchy: A Tool to Build Tools
National Bank of Canada wanted to custom build a tool that would integrate and easily share internal information and documentation with data scientists and business intelligence teams to support its big data efforts. Some solutions on the market were focused more on governance, rather than data pooling and aggregation — which was like “buying a plane to get a bicycle,” says Sebastien Beaulieu, senior manager of data, business intelligence and analytics strategy at the bank.

Enter Cinchy. Cinchy’s customization and flexibility gave various internal stakeholders a way to collaborate within the bank’s ecosystem to build the right tool. Beaulieu calls it “play dough.”

The Cinchy tool sits on top of the bank’s SQL (Structured Query Language) server. Implementation took six months, but Beaulieu says it could’ve taken as little as three.

While he wasn’t authorized to share specific performance metrics, he says there has been “tremendous value” from having a centralized, federated data lake. It used to take up to 200 days to integrate a new data source into the bank’s previous Oracle tool; with Cinchy, it now takes 10 days.

Empyrean Solutions: Modeling Everything
Pinnacle Financial Partners was preparing to grow over $10 billion in assets and knew its existing asset-liability management (ALM) platform wouldn’t meet the heightened regulatory expectations that come with crossing the threshold.

It learned about Empyrean Solutions from the recommendation of several larger peers. Today, the $29 billion bank uses the modeling platform for “just about everything,” says Brian Gilbert, the bank’s asset liability manager.

The Nashville, Tennessee-based bank uses Empyrean to model ALM, credit, funds transfer pricing, non-maturity deposits, contingency liquidity and net interest margin forecast. The bank is also using it to calculate its allowance under the current expected credit loss model, or CECL. Running 17 scenarios on the platform now only takes 17 minutes, he says. Such pulls can take hours to run at other banks; the time saved allows him to focus on the inputs in each situation and create multiple scenarios.

Gilbert couldn’t devote continuous full days to the implementation but estimated that a dedicated team could do it in a week. The platform doesn’t connect to a bank’s core; instead, users load data and financial information directly into it. These data sets can then be made available for any modeling, reducing the amount of raw extracts that Gilbert needs to perform.

“It’s hard for me to say how many more people we would need to have on my team if we didn’t have Empyrean. We would need more than four people to run interest rate forecasting,” he says.

INETCO Systems Limited: ATM Oversight
Boeing Employees’ Credit Union, or BECU, relies on its fleet of ATMs to deliver a large percentage of its financial services to members. It’s not unusual for one of its ATMs to have 14,000 transactions a month. So it’s a big deal when one goes offline and needs to be repaired or serviced, says Shirley Taylor, digital channel manager at the Tukwila, Washington-based credit union.

The $22 billion credit union wasn’t finding out about ATM outages or errors in a reliable or timely manner, in part because machines would occasionally go offline. Its ATM vendor introduced the credit union to INETCO, an operations intelligence platform that monitors transactions.

The credit union liked that the platform created alerts when there was an interruption in an ATM’s normal activity or when transactions failed to process — signs of a problem. Those alerts mean that fewer ATMs stay out of commission, allowing members to use its machines instead of sending them elsewhere. The fintech also generates potential fraud alerts and helps Taylor generate regular usage reports much faster. What used to take four to six weeks and involved reaching out to three other colleagues for data now takes an hour, she says.

Changing bank operations is never easy. But Cinchy, Empyrean Solutions and INETCO Systems Limited show how bankers can improve their institutions, saving employees valuable time and reducing the number needed to conduct essential operations.

Practical AI Considerations for Community Banks

A common misconception among many community bankers is that it isn’t necessary to evaluate (or re-evaluate for some) their use of artificial intelligence – especially in the current market climate.

In reality, these technologies absolutely need a closer look. While the Covid-19 crisis and Paycheck Protection Program difficulties put a recent spotlight on outdated financial technology, slow technology adoption is a long-standing issue that is exacerbating many concerning industry trends.

Over the last decade, community banks have faced massive disruption and consolidation — a progression that is likely to continue. It’s imperative that bank executives take a clear-eyed look at how advanced technologies such as AI can support their business objectives and make them more competitive, while gaining a better understanding of the requirements and risks at play.

Incorporating AI to Elevate Existing Business Processes
This may seem like a contrarian view, but banks do not need a specific, stand-alone AI strategy. The value of AI is its ability to improve upon existing structures and processes. Leadership teams need to be involved in the development process to identify opportunities where AI can tangibly drive business objectives, and manage expectations around the resources necessary to get the project up and running.

For example, community banks should review how AI can automate efficiencies into their existing compliance processes — particularly in the areas of anti-money laundering and Bank Secrecy Act compliance. This application of AI can free up manpower, reduces error rates and help banks make informed decisions while better serving their customers.

It’s necessary to have a strong link between a bank’s digital transformation program and AI program. When properly incorporated, AI helps community financial institutions better meet rising customer expectations and close the gap with large financial institutions that have heavily invested in their digital experiences.

Practical Steps for Incorporating AI
Once a bank decides the best path forward for implementing AI, there are a few technical and organizational steps to keep in mind:

Minimizing Technical Debt and “Dirty Data”: AI requires vast amounts of data to function. “Dirty data,” or information containing errors, is a real possibility. Additionally, developers regularly make trade-offs between speed and quality to keep projects moving, which can result in greater vulnerability to crashes. Managing these deficiencies, “or technical debt,” is crucial to the success of any AI solution. One way to minimize technical debt is to ensure that both the quantity and quality of data taken in by an AI system are carefully monitored. Organizations should also be highly intentional about the data they collect.More isn’t always better.

Minimizing technical debt and dirty data is also key to a smooth digital transformation process. Engineers can add value through new and competitive features rather than spending time and energy addressing errors — or worse, scrapping the existing infrastructure altogether.

Security & Risk Management: Security and risk management needs to be top-of-mind for community bankers any time they are looking to deploy new technologies, including leveraging AI. Most AI technologies are built by third-party vendors rather than in-house. Integrations can and likely will create vulnerabilities. To ensure security and risk management are built into your bank’s operating processes and remain of the highest priority, chief security officers should report directly to the CEO.

Managing risks that arise within AI systems is also crucial to avoid any interruptions. Effective risk management ties back to knowing exactly how and why changes affect the bank’s system. One common challenge is the accidental misuse of sensitive data or data being mistakenly revealed. Access to data should be tightly controlled by your organization.

Ongoing communication with employees is important since they are the front line when it comes to spotting potential issues. The root cause of any errors detected should be clearly tracked and understood so banks can make adjustments to the model and retrain the team as needed.

Resource Management: An O’Reilly Media survey from 2018 found that company culture was the leading impediment to AI adoption in the financial services sector. To address this, leaders should listen to and educate employees within each department as the company explores new applications. Having a robust change management program — not just for AI but for any digital transformation journey — is absolutely critical to success. Ongoing education around AI efforts will help garner support for future initiatives and empower employees to take a proactive role in the success of current projects.

At a glance, implementing AI technologies may seem daunting, but adopting a wait-and-see approach could prove detrimental — particularly for community banks. Smaller banks need to use every tool in their toolkit to survive in a consolidating market. AI poses a huge opportunity for community banks to become more innovative, competitive and prosperous.

Preparing to Be There for Your Community

The fallout from COVID-19 will likely take some time to stabilize. The personal and social costs are already significant, and neither is independent of economic and business disruptions.

Especially impacted are the businesses on Main Streets everywhere that are served by community banks. Community banks will be essential to any recovery, so it is important that they take steps now to ensure they’re positioned to make a difference.

The Challenge Of A “New Normal”
Financial markets were in “price discovery” mode this spring, but that phase is unlikely to last for long. If Treasury rates rise from their current levels, banks are likely to do well with their traditional models. But if they remain low, and spreads eventually stabilize to 2019 levels, nearly every institution will encounter pressure that could undermine their efforts to be a catalyst for Main Street’s recovery.

Bank Director’s recent piece “Uncharted Territory” warned that the experience of past financial crises could mislead bankers into complacency. Last time, dramatic reductions in funding costs boosted net interest margins, which helped banks offset dramatically higher loan losses. The difference today is that funding costs are already very low — leaving little room for similar reductions.

Consider asset yields. Even without significant credit charge-offs, community bank profitability could face headwinds. Community banks entered 2020 with plenty of fuel to support their thriving Main Streets. Their balance sheets had been established for a Treasury rate environment that was 100 basis points higher than today’s. If rates settle here for the next couple of years and existing assets get replaced at “new normal” levels, yields will fall and net interest margins, or NIMs, could take a hit.

Banks could have trouble “being there” for their communities.

Where do the current assets on banks’ balance sheets come from? They were added in 2018, 2019 and the first quarter of 2020. If we assume a fixed rate loan portfolio yields somewhere around 300 basis points over the 5-year swap rate at closing (which averaged about 1.75% over 2019), and floating rates loans yield somewhere around 50 basis points over prime day to day, we can estimate banks’ first quarter loan yields at perhaps 4.75% fixed-rate and 5.25% prime-based.

Prime-based yields have already dropped for the second quarter and beyond: They are now earning 3.75%. Fixed-rate loans continue to earn something like 4.75%, for now.

Banks that can quickly reduce funding costs might, in fact, see a short-term bump in net interest margins. If they can stave off provision expenses, this might even translate into a bump in profitability. But it will not last.

If Treasury rates remain at these historic lows and spreads normalize to 2019 levels, current balance sheets will decay. Adjustments today, before this happens, are the only real defense.

Banks’ fixed rate loans will mature or refinance at much lower rates — around 3.50%, according to our assumptions. Eventually, banks that enjoyed a 3.50% NIM in 2019 will be looking at sustained NIMs closer to 2.50%, even after accounting for reduced funding costs, if they take no corrective steps today. It will be difficult for these banks to “be there” for Main Street, especially if provision expenses begin to emerge.

Every community bank should immediately assess its NIM decay path. How long will it take to get to the bottom? This knowledge will help scale and motivate immediate corrective actions.

For most banks, this is probably a downslope of 18 to 30 months. For some, it will happen much more rapidly. The data required may be in asset and liability management reports. Note that if your bank is using year-end reports, the intervening rate moves mean that the data in the “100 basis points shock” scenario from that report would represent the current rates unchanged “baseline.” Reports that do not run income simulations for four or more years will also likely miss the full NIM contraction, which must be analyzed to incorporate full asset turnover and beyond.

Times are hectic for community banks, but in many cases commissioning a stand-alone analysis, above and beyond standard asset-liability compliance requirements, is warranted.

Then What?
The purpose of analyzing a bank’s NIM timeline is not to determine when to start taking action, but to correctly size and scope the immediate action.  All the levers on the balance sheet— assets, liabilities, maybe even derivatives — must be coordinated to defend long-term NIM and the bank’s ability to assist in Main Street’s recovery.

The Small Business Administration’s Paycheck Protection Program lending is fully aligned with the community bank mission, but it is short term. Banks must also plan for sustainable net interest income for three, four and five years into the future, and that planning and execution should take place now. The devised NIM defense strategy should be subjected to the same NIM decay analysis applied to the current balance sheet; if it’s insufficient, executives should consider even more significant adjustments for immediate action.

The economic environment is out of bankers’ control. Their responses are not, but these require action in advance. Banks can — and should — conduct a disciplined, diagnostic analysis of their NIM decay path and then correct it. This interest rate environment could be with us for some time to come.

Data In The Best, And Worst, Of Times

Helping their community and delivering personalized service is the foundational differentiation of every community bank. Now more than ever, customers expect that their community bank understands them and is looking out for their best interests.

Customers are communicating with their banks every day through their transactions — regardless if they are mobile, in person or online, each interaction tells a story. Are you listening to what they’re telling you? Whether your bank is navigating through today’s COVID-19 crisis or operating in the best of times, data will be key to success today and in the future.

Business intelligence to navigate daily operations is hard to come by on a good day, much less when things are in a pandemic disarray. Many bankers are working remotely for the first time and find themselves crippled by the lack of access to actionable data. A robust data analytics tool enables employees at all levels to efficiently access the massive amounts of customer, market, product, trend and service data that resides in your core and ancillary systems. Actionable data analytics can empower front-line bankers and risk managers to make data-driven decisions by improving and leveraging insight into the components that affect loan, deposit and revenue growth. Additionally, these tools often do the heavy lifting, resulting in organizational efficiencies that allow your bankers and executives to focus on strategic decision-making — not managing cumbersome data and reporting processes.

A tool that aggregates transformative data points from various siloed systems and makes them readily available and easy to interpret allows your management team to be better prepared to proactively manage and anticipate the potential impact of a crisis. This positions your bank to offer products and services that your customers need, when they need them.

But most community banks have not implemented a data analytics solution and as such, they  must consider how to manually generate the information needed to monitor and track customer behaviors to assist them in navigating this crisis. Below are a few potential early warning indicators to monitor and track as your bank navigates the current coronavirus crisis so you can proactively reach out to customers:

  • Overdrafts, particularly for customers who have never overdrawn.
  • Missing regular ACH deposits.
  • Past due loans, particularly customers who are past due for the first time.
  • Line of credit advances maxing out.
  • Lines of credit that cannot meet the 30-day pay-down requirement.
  • Declining deposit balances.
  • Large deposit withdrawals.
  • Businesses in industries that are suffering the most.

If your community bank is one of the many that are proactively assisting customers during this pandemic, make sure you are tracking data in a manner that allows you to clearly understand the impact this crisis is having on your bank and share with your community how you were able to help your customers during this critical time. Some examples include:

  • Paycheck Protection Loan Program details: number of applications received, processed and funded; amount forgiven; cost of participating for the bank; customer versus non-customer participation, impact on lending team, performance.
  • Customer assistance with online banking: How did you help those who are unfamiliar with online banking services? How many did you assist?
  • Loan modifications, including extensions, deferments, payment relief, interest-only payments and payment deferrals.
  • Waived fees and late charges.
  • Emergency line of credits for small business customers.

Having easy access to critical customer information and insights has never been more important than it is today, with the move to remote work for many bankers and rapidly changing customer behaviors due to the economic shutdown. Customers are making tough choices; with the right data in your bankers’ hands, you will have the ability to step up and serve them in ways that may just make them customers for life.