How AI Tackles Three Categories of Risk in Financial Services
AI is just as valuable to financial services companies as any other when it comes to mitigating cybersecurity risk, and it can also be helpful in managing compliance with cybersecurity frameworks and regulations.
Beyond that, though, are three categories of financial risk that AI can help organizations detect and reduce: financial fraud, bad loans and bad investments.
AI can detect fraudulent transactions and claims. Banks, credit card companies and others in the industry have been using AI to help detect and eliminate fraud for many years, and the technology continues to improve. AI systems are adept at learning customers’ behavior patterns and then applying what they know during individual transactions.
For instance, suppose a customer known to live in Pittsburgh makes a transaction at a grocery store she has shopped at previously — no problem there. Now, suppose a subsequent transaction is presented at a St. Louis gas station. That transaction would probably be flagged by the card issuer’s fraud detection system.
That’s a pretty obvious example, but sophisticated AI-powered fraud detection systems can use many factors to detect potentially fraudulent transactions, including geography, time of day, type of transaction, number of transactions in a short period and more.
AI is also being used to detect other kinds of financial fraud. For instance, according to the Coalition Against Insurance Fraud, about 56 percent of insurance companies already use AI systems to detect potentially fraudulent claims, which cost the industry nearly $309 billion a year.
LEARN MORE: In 2024 AI will lead digital transformation efforts for financial services.
AI can help lenders make better decisions. Banks and other lenders can get a better idea of how well a potential borrower fits within their credit profile by using well-trained AI systems. Modern lending isn’t conducted the way it was a generation ago, when potential borrowers would wear fancy clothes and present themselves to an agent at the savings and loan.
Today, the process is digital and largely commoditized. Applicants fill out forms online, often through a so-called affiliate marketing channel such as LendingTree. Lenders then compete against each other for an applicant’s business. “Affiliate channels are particularly competitive, making it critical for lenders to have more accurate predictors of credit risk,” Garrett Laird, director of decision science for the fintech company Amount, explained in a recent article. “Simple credit policies relying primarily on general credit scores like FICO will not capture the variance present within each FICO quality band.”
Lenders “have to offer the features the would-be borrower wants and the applicant, on the other hand, has to fit the institution’s credit profile,” he added.
Al can help financial services companies make better investments. The emergence of a category of financial services companies whose advice and investment decisions on behalf of clients are driven largely by AI models rather than live advisors began some years ago. “Robo-advisors” gather information about clients’ financial goals and risk tolerance, then manage their money accordingly.
DIG DEEPER: How AI is used in the financial world today.
But, in fact, all financial services companies that make investment decisions — on their own behalf or their clients’ — are finding AI an irresistible tool. A survey of financial advisors by Accenture found that 98 percent of them “believe that AI is transforming how advice is created for, delivered to and consumed by clients, and 97 percent believe that AI can help grow their book of business organically by more than 20 percent,” according to a press release.
The financial services industry still has a ways to go when it comes to actually implementing AI, though. A separate Accenture study “found that the most AI-mature companies achieved 50 percent higher revenue growth than their peers and that the capital markets and banking industry, which includes wealth management, had the lowest AI maturity score out of 17 industries analyzed,” a press release notes.
RELATED: From advanced computing to predictive analytics, checkout CDW’s AI offerings.
AI Is Just as Prone to Bias as Humans Are
Within each of these areas of AI opportunity, however, comes an important caveat: Artificial intelligence still can’t replace human attention entirely, and the technology can actually exacerbate humans’ unfortunate penchant for bias.
AI models are built and trained by humans. As a result, they can have the same blind spots that people do. Take lending, for example: An AI model can be a boon to a financial services company seeking insights about whom to lend money and at what interest rates, but if the model includes discriminatory factors, then it will screen people out, or set unfair loan pricing levels based on those factors.
The same can happen with fraud detection. When a legitimate transaction is flagged as potentially fraudulent, it creates an inconvenience for the customer. So, companies need to be careful about, say, fraud detection models that flag certain locations as being prone to fraud, which could create a chronically poor customer-service experience for everyone who lives in that area.
The point is that AI is only as good as its training. For a more detailed discussion on how organizations can best make use of AI systems, and on what time frame, feel free to reach out.
This article is part of BizTech's EquITy blog series.
