By Adam Devine, WorkFusion
Artificial Intelligence (AI) technology, once relegated to science fiction, is a profitable reality at the core of many businesses. Banking is no exception. With margin challenges and increased customer demand for improved services, many forward-thinking banks have already made AI a part of their strategy, from front to back office operations.
While AI has endless application in banking, an obvious area that suffers from manual effort and swelling cost is back office functions. Many rote and time-consuming tasks like employee and customer onboarding, KYC, data migration, and P2P can be performed exponentially more effectively and efficiently with both robotic process automation (RPA) and AI, rocketing a bank from the analog times to the new digital era.
What does AI actually do for banks, and how does it complement RPA?
There’s an equal amount of talk and confusion within banks about what AI actually does. Simply put, AI gives machines the ability to learn, reason, and understand. It uses historical data and real-time human action to train algorithms to do work the same way a person would – only faster and without errors. AI won’t necessarily replace human intelligence in many functions. It will augment it.
RPA has swept through most large banking operations over the past two years, but that doesn’t mean its utility is understood. Whereas AI uses a history of actions to predict an action that requires some degree of judgment, RPA uses defined programming to execute an action that requires only rules. RPA is ideal when tasks within a process have no variation, like entering user credentials into Oracle or moving structured data from one system to another. AI and RPA fulfill different and complementary roles within a process and together can digitize up to 90% of many core processes. This pairing of AI and RPA is often referred to as intelligent automation, or IA.
Modern banks use IA to provide an Iron Man suit to customer-facing and back office teams alike, allowing people to focus on surprising and delighting customers and handling complex exceptions that automation, or “bots”, can’t tackle. Four arenas of many where IA will have a big impact are compliance, back office transaction processes, call center operations, and HR. Below, I’ll explore it’s practical application for compliance.
How does AI digitize compliance?
Regulatory compliance is a time and cost vortex for every bank. Much of the work is manual and paper-intensive. This is an area where IA delivers.
One example where IA can improve compliance is separating false positives from true compliance violations. It also aids compliance officers in reading and parsing through lengthy regulations and makes regulatory exams and reporting much less strenuous for banks. IA can even prevent fines and preserve reputations in trade surveillance by using natural language processes (NLP) to detect rogue behavior in traders’ emails and chats.
Smart banks are already moving forward in these areas by developing internal capabilities and leveraging best-in-breed products. For example, Credit Suisse is developing AI software that monitors employees for rogue behavior. The bank is initially focused on detecting unauthorized trading. Over time its technology will monitor all employee behavior for breaches of conduct rules.
This not only helps banks operate more efficiently but also avoids costly and potentially criminal behavior. For instance, IA can identify fake accounts by determining if they were opened with the same email address or from the same IP location. It can also be used to pre-emptively identify email address names that have a high likelihood of being fraudulent. Wells Fargo would have avoided its fraudulent account scandal had it been equipped with this capability.
Imagine the opportunities and capacity within banking operations once the process of mapping and identifying regulatory compliance is automated. Rather than being a siloed operation, financial institutions will be able to make regulation a central part of business decisions. Regulatory requirements can be linked to controls in a library that has costs associated with each control, which makes it possible to look at and compare the cost of regulatory compliance in each business unit.
Fraud detection is another area where IA can not only automate operations, but also improve them at the same time. Fraud detection, like compliance, is a time-intensive process requiring many employees to parse through large amounts of data to detect patterns and anomalies. Smart financial institutions are leveraging machine learning, a core capability within intelligent automation, to help.
According to a recent article, banks can use IA technology to conduct an ongoing review of account activity patterns and flag aberrations for further review. Over the last decade, IA has not only significantly improved the monitoring process, but is now able to respond in real-time to potential fraud. This is critical not only to make operations more efficient, but also to protect against a rising amount of fraud and cybercrime. Financial institutions can use IA to evaluate transactions and millions of data points in real-time and dynamically shift between response models as threat levels change.
How can banks use AI as both defense and offense?
A bank today is like a peach: regulatory compliance is the pit, and profitable business units are the juicy flesh. Born-digital competitors are eating away at the flesh from the outside, and regulatory compliance is eating away at the flesh from the inside. Without swift and assertive operational transformation, banks will be left only with the pit. IA is a way to stop profit erosion from both within and without by giving these venerable institutions the agility and elasticity of born-digital businesses. The world’s smartest, most nimble banks will use IA as both armor and propulsion, retaining and winning more customers, leaning and digitizing operations, and developing new ways to provide value in a world where people and machines work together.