4 Transformative Applications for AI & Machine Learning

Across the world of fintech, artificial intelligence (AI) and machine learning are resoundingly touted as the Next Big Thing.  In fact, according to PwC, AI companies received more funding in 2017 than ever before, with 52 percent of financial services decision makers surveyed stating that they are making substantial investments” in the technology, and 72 percent referring to AI as the “business advantage of the future.” Accenture Research, cited by CIO Magazine, projects that companies in the financial services sector that embrace AI could improve profitability by an average of 31 percent by 2035. As AI and machine learning are operationalized by credit unions, their potential is almost too vast to predict. To truly drive benefits in the near term, financial institutions must understand how to first apply these tools to realize the greatest gains. Two Distinct Technologies According to Fotis Konstantinidis, Senior Vice President, Fraud Products for CO-OP Financial Services, the terms AI and machine learning are frequently used interchangeably, though they are two distinct technologies. “AI as a concept envisions a machine that thinks like a human,” Konstantinidis says. “It is an umbrella term that encompasses any range of applications, from perception of the environment through sensors, reasoning using rules, visual recognition, voice recognition and machine learning. Machine learning refers to technologies that teach computers to learn — and then apply this learning going forward.” Initially, says Konstantinidis, credit unions should focus AI strategies and investments on areas that promise to improve the member experience the most. “Start with the member need, determine the ideal outcome, and then apply AI and machine learning platforms to solve problems — whether you are looking for a way to expand member service or streamline fraud detection,” he says. To that end, he has identified four areas where machine learning can dramatically — and quickly — improve outcomes. 1. Enhanced Member Insights “The more we understand our members, the better we can serve them,” said Konstantinidis. “Credit unions and banks have all this data, but they don’t use it to deliver real value to the consumer. For example, everyone sees the same credit card or mortgage promotion. When products are not personalized, credit unions miss opportunities to build engagement and loyalty.” AI technologies today, he adds, allow credit unions to access not only their internal data but also external data from social media feeds and other sources to understand who their members are and then offer them products and services tailored to them. “Let’s say a member just had a third child,” he says. “Offering a great rate on an auto loan might serve this growing family well. And, credit unions should automatically know when members are traveling so they don’t bombard them with messages about what they perceive to be suspicious activity. These member interactions are not well contextualized today, but they can be given the technology and data available.” 2. Efficient, Omnichannel Boarding Traditionally, Konstantinidis notes, onboarding new members has required in-branch visits and a plethora of paperwork to authenticate individuals and secure the required approvals. “Today, there are many different types of devices in the hands of consumers that we didn’t have in the past,” he says. “Credit unions need to be able to onboard members digitally in the age of Big Data using tools like mobile apps and social network accounts. AI can assist tremendously in authenticating a digital identity, which helps simplify processes and gets rid of the paper trail.” 3. Increased Operational Efficiency To streamline their operations, Konstantinidis says, credit unions should consider deploying bots: “Bots perform consumer-facing tasks very efficiently. They can answer questions and deliver services to members 24/7, expediting processes and reducing demands on call centers.” Bots excel at back-end tasks too — and can be deployed to interact with systems in ways that previously required human skill. “Using robotic process automation (RPA), credit unions can automate many repetitive tasks on the operational side and, for example, deliver a risk score for a new member immediately,” he says. 4. Improved Risk Management When it comes to risk management, financial institutions care about two main things, “fraud losses and false positives," Konstantinidis affirms. "Machine learning is at the point today where risk management can be significantly improved. Every transaction that goes through the model is ingested and understood by the machine instantaneously, versus the days or weeks involved in human analysis." He adds that machines can employ deep learning as well for greater precision, and follow “a predictive approach versus a reactive one — seeing transactions from beginning to end and understanding when they constitute uncharacteristic behavior.” CO-OP’s AI Initiative Meet COOPER, an advanced data-driven platform designed to detect and fight fraud faster than ever before. CO-OP President and CEO Todd Clark introduced the new platform at the THINK 18 Conference in Arizona  “COOPER is a major piece of our strategy to bring greater security across our products and services, while providing the most seamless experience to the member,” Clark says. “The machine learning power of COOPER will allow us to constantly improve upon the fight against fraud by enabling the understanding of huge amounts of data and detecting complex patterns rapidly.” According to CO-OP data scientist and Senior Vice President, Fraud Products Fotis Konstantinidis, credit unions are increasingly determined to put AI to work in the fight against fraud – and beyond. “This is one of the main reasons CO-OP has made machine learning for fraud a priority,” he says. “Credit unions need the ability to leverage this technology, and our capacity to integrate machine learning across card, account and call-center networks allows us to make the best use of its speed and computational power.” He continues, “We read a lot of articles today in which AI is portrayed as the panacea to all our difficulties. Apowerful as these technologies are, they will not solve all our problems magically. We still need human intervention, and we will ultimately achieve the minimal amount of human interaction, where the machines adapt extremely fast to human input. It is up to us to guide the machines and establish our objectives for them. Only then will we achieve the intended results.” As COOPER ingests more data and continues to learn and adapt, more use cases and eventually AI will help evolve fraud detection on a constant basis, provide more detailed and predictive business intelligence, and improve and personalize the member experience. For credit unions looking to add AI and machine learning to their fraud-fighting arsenal, COOPER is a welcome platform for change. COOPER is joining your team in the fight against fraud. To learn more, visit co-opfs.org/cooper.