Insurance companies don’t always inspire a great deal of sympathy; after all, who enjoys paying their car insurance premium? Who hasn’t felt a bit like they’re constantly paying for something they’re afraid to use for fear that their rates will go even higher?
But insurance companies are engaged in a war. And believe it or not, you’re not the enemy. You’re one of the victims.
The war is against criminals—from large, sophisticated enterprises involving doctors, insurance agents, and other professionals, to backyard get-rich-quick schemers—who are constantly devising new ways to fraudulently collect on insurance policies. They’re costing all of us our hard-fought money, driving up insurance rates for innocent bystanders while stripping away a sizable portion of insurance companies’ earnings.
Hanzo is proud to be part of the war effort, using our web-crawling technology to fight back.
How We Got Here
We didn’t start out as insurance fraud investigators! Hanzo got its start as a two-man technology startup inspired by the web archival project of the British Library. Our first big break came when U.S. companies contacted us because they needed help retaining records of their web-based content for regulatory compliance. From there, we kept building our web-crawling and archival technology, and we started getting new customers asking us to help them preserve online evidence for ediscovery.
Over the years, web-based content has only gotten more complex and interactive, from mouse-over animations to drop-down menus and chatbots. Hanzo has grown right along with the content, specializing in complex, dynamic, unstructured data identification, collection, and preservation. And our methods have been court-ready from the start: admissible, authenticated, and based on a solid, demonstrable evidentiary foundation.
That’s where we were when we started getting a new kind of customer inquiry: insurance companies asking us to help them spot patterns in online content that might indicate fraudulent activity.
The War on Insurance Fraud
You may not realize it, but insurance fraud is a massive problem in the U.S. and beyond. The Coalition Against Insurance Fraud estimates that no less than $80 billion is stolen every year, just in the U.S., through fraudulent activities. Some claims started out legitimate before the customer decided to inflate the value of their loss; other more dedicated criminals created damage where there wasn’t any. These cases run the gamut from individuals who decided to burn down the houses they couldn’t sell to serious criminal conspiracies, like the roofing contractors who damaged roofs they were inspecting so they could cash in on insurance payouts. (Those class acts were apprehended when video surveillance showed them tearing up tiles to mimic storm damage.) And that’s one of the smaller schemes; some health insurance cons can extend for years, like the sleep study clinic whose founder stole at least $83 million from Medicare and private insurance companies for entirely unnecessary sleep studies.
And it’s not just insurance customers and associated professionals stealing money; according to the National Association of Insurance Commissioners, dishonest insurance agents and entirely fake companies collect premiums from unwitting consumers without ever purchasing insurance for them. In a single healthcare fraud investigation, the government found nearly 150 such fake insurers that had swindled 200,000 people, leaving them holding the bag for over $250 million worth of unpaid claims.
Perhaps insurance fraud is so common because it feels “victimless” and seems like an “easy” crime to commit. All it takes, in many cases, is a little looseness with the truth, a minor exaggeration of the extent of damages caused by an accident or a storm. Besides, haven’t these poor saps been paying for insurance for years with nothing to show for it? The FBI attributes much of the problem to the size of the insurance industry, opening up “more opportunities and bigger incentives” for criminal activity.
But insurance fraud is a long way from victimless. It results in increased premiums for all of us, to the tune of $400 to $700 a year. And fraud siphons off up to 10 percent of insurance companies’ revenue.
Fortunately, insurance companies have found a way to bring the battle to these criminals.
Using AI to Fight Insurance Fraud …
AI, or artificial intelligence, refers to technology that allows computers to simulate human intelligence. One common form of AI, machine learning, feeds vast quantities of data to computers, which then process that information while making rules about what data means. Based on feedback from human programmers, these AI systems learn to think for themselves, extrapolating from rules and information they’ve been given to create their own entirely new rules—which no humans are privy to. This process points out the symbiotic collaboration that AI necessitates between people and machines: humans provide the feedback that helps machines learn, after which machines can point humans toward helpful pieces of data and overarching patterns that would be impossible for any human to spot.
Such AI systems have been used to churn through enormous datasets about insurance claims, customers, fact patterns, involved doctors and healthcare facilities, geographic locations, and more. Because they can see and process so much data so rapidly, these computers are able to identify suspicious patterns of information that human investigators would miss. Even IBM is in on the action, using its Watson AI to provide data analytics for a Counter-Fraud Management program.
… By Crawling the Web for Information
We developed our web-crawling technology to gather and preserve information for regulatory compliance and ediscovery, but that doesn’t mean it’s only useful for those limited purposes. Our archival experience taught us how to do effective, comprehensive web crawls across all kinds of social media platforms and dynamic, interactive websites. We can now use it to look everywhere on the web for everything there is to find about people who’ve made, or been involved in, insurance claims.
With our web-crawling and native-format web capture technology, Hanzo is helping insurance companies investigate potentially fraudulent claims. We can locate all of the social media sites a claimant uses and verify their activities. We can dig around the entire internet to identify suspicious patterns or incriminating evidence—and our evidence will be court-ready, because that’s how we built our software.
Is someone who claimed to be injured in a car accident skiing in the Rockies? Going out with friends for drinks and dancing? Using their workers’ compensation time to get some home-improvement projects done? Or maybe someone who claimed to be the victim of theft or property damage is, miraculously, listing on eBay the same items they claimed to have lost. Are people who have made similar claims connected in some subtle way that can be detected through their online activity? (You’d think that people who were trying to pull off a heist wouldn’t post about their activities on social media, but you’d be surprised. Perhaps they enjoy the thrill of getting away with something and find it less thrilling if no one else knows about it!)
Not all of the evidence we find is so obvious: sometimes the smoking gun is hidden away in location metadata on a photo or in a time-and-date stamp that contradicts a claim. Sometimes our technology finds criminal histories or a history of other suspicious claims that tells our clients where to focus their investigative efforts.
Because all this emphasis on technology isn’t to say that computers can detect and establish fraudulent activity all on their own. It takes a village: boots on the ground—real people investigating suspected fraud using tried-and-true surveillance techniques and old-fashioned detective work—and smart computers behind the scenes, identifying potential criminal schemes and bad actors that the human component can focus on. Computers can crunch through volumes of data to spot patterns, but humans can apply intuition, experience, and creative thinking to get to the bottom of a situation.
Hanzo is honored to help in this war. If you see another way that our web-crawling technology could aid in investigations or branch out into yet another industry vertical, please get in touch. Our growth has always been driven by our customers’ needs, and we love where it’s taken us so far.