CoventBridge Group’s typical undercover surveillance case takes place in broad daylight, often in a medical center’s parking lot, so that the investigator can see a potential suspect clearly. Does he or she put on a neck brace only after getting out of the car? Does the limp start just as they reach the building lobby?
In addition to employing traditional steps such as interviewing neighbors or coworkers who may have heard a subject boast about crashing their car into a post on purpose, Jason Zurn, chief information officer, is helping the company enhance SmartPartner, its proprietary case management system with an innovative artificial intelligence (AI) solution. Using predictive analytics, SmartPartner aggregates a broad range of relevant information that includes all aspects of claim data and a review of claimant activity for “red flags” that indicate potentially fraudulent claims that need further investigation.
It’s all part of the world’s largest insurance fraud investigation company’s efforts to help insurance companies combat the estimated $80 billion in annual fraudulent claims. These claims cost the average US household an additional $200–$300 in premiums each year.
Finding Fraud in Volume
According to Zurn, part of the problem is the high number of claims. “Many claims adjusters handle hundreds of files at once, so by the time they gather enough evidence to suspect fraud, it’s been going on for a long time,” he says.
To help address the problem, CoventBridge has developed a unique approach to detect fraud within insurance claims without human intervention. The AI solution takes a proactive approach by comparing a client company’s claims data against measurable indicators. To do so, it scores the likelihood of fraud in each case and refers red-flagged claims to the insurer.
“The sooner we can identify the risk of fraud, perhaps as early as the underwriting process, the more we can help insurers save—and possibly even prevent the problem before it happens,” Zurn says.
The Inner Workings of Predictive Analytics
The predictive analytical engine works by aggregating a wide variety of data types. These can include cases without witnesses to an incident, a long delay before reporting an accident, filing a claim very shortly after a policy takes effect, or spotting claimants who change physicians frequently or may have ongoing financial issues. In isolation, none of these details are necessarily significant. However, in context with other objective indicators, they become more meaningful.
For each engagement, CoventBridge provides standard metrics it has found to be most effective. It also customizes the analytics to align with each client’s priorities.
One of Zurn’s challenges in providing the service is that many insurance companies still rely on legacy systems, some dating back to the mid-1990s. “SmartPartner is web-accessible. We have to provide compatibility with some very old equipment and still ensure the security of very sensitive information,” he says. “Despite that, technology issues have never stood in the way of providing a fully turnkey solution for our customers.”
The AI service eliminates the need for clients to invest in IT upgrades, additional manpower, or licensing fees. It has also been shown to deliver results that improve over time. That’s because Zurn and his team provide feedback to the system on the accuracy of previous results. This helps it “learn” to better identify red flag behaviors and subsequent fraudulent activity.
Keeping Information Safe and Clean
In addition to uncovering fraud earlier, CoventBridge scrubs client data. This removes all anomalies and discrepancies and delivers cleaner, more structured information back to them. Company services also enable insurers to attribute their fees to Allocated Loss Adjustment Expenses (ALAE). The accounting benefit helps more accurately determine reserves required for claim payouts and processing.
To maintain high levels of internal and external customer service, Zurn follows a straightforward motto: “If I can do it faster than delegating, then it’s better for clients and staff.” Even though he is on the leadership team, it isn’t uncommon for Zurn to take care of something like a simple help desk ticket.
“If what’s needed is a five-second password reset and that gets someone back on the system a half-hour faster, I’ll take care of it,” he says. “It’s not the kind of thing that highlights how smart you are. It makes life easier for the rest of the staff and makes you the kind of guy they want to keep working for.”
The approach also helps Zurn stay up to date with all current IT issues within the company. Zurn supplements that information with weekly meetings in which CoventBridge executives share pertinent concerns for their departments. This provides context and insight for the entire team.
Zurn uses that to his advantage. “I view everything as a problem to solve, a ‘solution opportunity,’” he says. “The more I know, the more I can keep shoveling coal into the engine so the train doesn’t slow down.”