Artificial Intelligence Technology Transforms Software Solutions

For many people not directly involved in the development of software for the insurance industry, any discussion of new technology can rapidly swell into a serious information overload. Compounding this problem is the oft-noted “snail’s pace” at which new technology has been embraced by the industry as a whole, particularly by carriers.

“The insurance industry does not openly [promote] the type of environment that grasps software technology,” said John D’Alusio, senior vice president of strategic relations and e-business for Irvine, Calif.-based HNC Insurance Solutions. “There’s a myriad of reasons why they’re not grasping the technological availability out there as some might think they would. A lot of that is an aversion to change.”

But the pace will, by necessity, pick up in the near future as the type of market conditions especially evident in the current workers’ comp arena compels companies to find new ways of becoming more profitable and more efficient.

Neural networks
“The best way to do that is through technological advances, such as neural networks,” D’Alusio said. He believes that once one major carrier makes the investment and begins to incorporate such advances, improvements in efficiency will make it imperative for other companies to follow suit in order to stay competitive and keep costs down.

In a nutshell, a neural network is artificial intelligence. That is, a neural net-based system is a highly sophisticated technology that involves the ability of the system to think and learn. Other companies in the neural net area include Blaze Software Inc. and Fair, Isaac and Co. Inc. The technology has already been successfully introduced in a number of areas.

On the claims processing and underwriting side, no rules-based decision engine “productized” (turned into a product) for workers’ comp has emerged in the market as yet. However, HNC Insurance Solutions is in the process of building one.

“The first thing you have to do for a neural network is build business decision rules that it fires against,” D’Alusio explained. “We have built the generic decision rules for workers’ compensation claims, and we expect to have this productized within the next 12 months where we could actually go and show our clients a working model and then further customize it to their requests.”

Boosting profits
Started during the first half of the 1980s, medical bill review has been estimated as now being a $200-million industry in workers’ comp alone. After the first couple of carriers that used it wound up paying less in their medical charges, using medical bill review became an industry imperative in order for carriers to remain competitive.

D’Alusio said that the product responsible for most of his company’s $83 million in sales a year is the medical bill review system COMPADVISOR™, a version of which exists with its Capstone neural network rules-based decision engine. “You can use it to refine and refine how you review medical bills to the point where perhaps 80 percent of all of them are being system-adjudicated without the need for any human or manual intervention,” he said.

Among the numerous companies involved in the delivery of medical bill review software is Ingenix, a segment of UnitedHealth Group, which recently made its PowerTrak Version 4.0 available to the property/casualty market.

A plethora of smaller, more regional companies are involved in the medical bill review software industry. But while bill review software is definitely big, no major technological advances have been made in the area in the recent past.

“It’s essentially been refinement of systems that have been previously built—now the push is to have it Internet-based,” D’Alusio said. “But it’s still the same requirement. You have to look at the treatment that was rendered, the injury, to make sure that there’s a causal connection and that the treatment being rendered is proper for the accident and injury that occurred, and that the billing is within the guidelines.”

Detecting comp fraud
Another major area of focus has been the development of tools to assist companies with workers’ comp fraud detection.

“Fraud and abuse within the insurance sector cost corporations, states, the federal government and individual taxpayers billions of dollars each year,” D’Alusio said. “The problem will continue until it is addressed with a combination of ‘smart’ anti-fraud software solutions, aggressive special investigative units, and cooperation from district attorneys’ offices.”

Most of the technological advances in this area have been based on discovering claimant fraud. Historically the claims adjuster responsible for a file would also be responsible for detecting fraud, which could be a hit or miss affair due to the different levels of training and experience of individual claims adjusters in recognizing fraudulent representation.

“Our software will rate a case from zero to 1,000, zero meaning there’s absolutely no possibility of fraud, and 1,000 meaning that it’s highly possible that there’s fraud involved,” D’Alusio explained.

Neural network technology is also used here, in this case to study statistical patterns and determine their effect. “It’s not just a probability system,” D’Alusio continued. “Carriers that use it set a particular figure—it could be 400, 600 or 750—where every case reaching that number is directed to a special investigative unit for an intense anti-fraud investigation.”

To help combat employer fraud, HNC has also developed anti-employer fraud software called VeriComp™ Employer. “On workers’ compensation, premium rates develop based on class of employees,” D’Alusio said. “A steelworker has a much higher workers’ compensation rate than a white-collar worker. …So there is a possibility for fraud there, too.”

VeriComp™ Employer can alert carriers as to which employers may be engaged in fraudulent submissions. However, D’Alusio noted that the tool has been somewhat lightly used due to a perception on the part of some agents that the product might alienate the employers using those agents as their representatives.

Referring to HNC’s VeriComp™ Claimant product, D’Alusio said, “We found that the anti-claimant fraud technology has had much more success initially in the marketplace.”

Another major entity involved in anti-fraud technology software is Insurance Services Office Inc., which has been making major strides towards positioning itself as a leading provider of electronic services to the property/casualty industry. Among the many products developed by ISO are ClaimSearch®, an all-claims database and, more recently, ViewLink Manager™, which the company describes as “a PC software tool that helps identify suspicious claims and reduce delays in paying legitimate claims.”

In addition, last November, ISO joined Capital Z Financial Services Fund II, a global private equity fund, in an investment in eStellarNet, a San Francisco Bay area company dedicated to electronic connectivity between medical providers and carriers.

In summing up his company’s various products developed for the insurance industry, D’Alusio said, “The way we approach it is that the value play is the back-end of the organization, on the claims side, not the front-end distribution channel. …Overwhelmingly, most of the money goes out to pay claims and to pay your claims adjusters.”

Topics Fraud Claims InsurTech Workers' Compensation Tech Data Driven Artificial Intelligence Market