Skip to main content

How Artificial Intelligence Is Changing Workers’ Compensation

Artificial intelligence (AI) is proving to be a valuable tool in managing workers’ compensation claims, particularly regarding litigation avoidance.

Applying the technology to workers’ compensation claims can also produce other benefits [1], including a better understanding of the source of claims, easier selection of a best-in-class medical provider for a claim, and quicker identification of potentially catastrophic claims for early intervention.

As with other technologies, companies that embrace the innovation’s potential earliest are in a position to achieve the greatest benefits because they gain an advantage over competitors.

 

A Relative Wealth of Data to Be Analyzed

Compared to many other areas of risk, workers’ compensation has the advantage of associated state requirements, which result in the collection of a large amount of data — information that goes beyond simple details of the injury.

On top of using AI to assess that structured data, natural language processing can be applied to regular communications between a claims-adjusting team and an injured worker. Those evaluations can provide significant insights into the employee’s sentiment.

Assessing that sentiment can provide a view into the general rapport between the injured worker and the claims administrator. Such information can be useful in predicting litigation risk. If the sentiment is judged to be negative, there’s a greater risk of litigation; if it’s positive, the risk of litigation is less. Understanding that risk can help guide the employer’s approach to the claim.

Over time, the machine-learning elements built into the AI result in continuous refinement of the models. Each refinement might be small, but they accumulate over time — the longer the models are in place, the more accurate they’re going to be. As the models improve, the value of the data grows, as does the technology’s ability to predict the risks associated with workers’ compensation claims.

Ultimately, being in a position to analyze all the data points available, both structured and unstructured, can help employers better understand injury trends. AI is even helping identify workers’ compensation fraud. Those results, in turn, can reduce loss costs and total cost of risk.

 

Sorting Out the Pandemic’s Impact

As the COVID-19 pandemic caused many employees to have to work remotely, employers often lost control of the ergonomics of those employees’ workplaces, setting the stage for potential workers’ compensation claims.

While a wave of claims from workers setting up shop on dining room tables or living room couches has yet to emerge, it likely will arrive. As those claims appear, AI is expected to help employers manage the risks.

AI will help employers understand which cases might be serious by predicting the relative complexity of the case based on its attributes. The technology will consider not only diagnostic codes but factors such as where the employee lives in determining whether the claim might present a heightened risk.

The detailed analysis of those claims will also help employers in regard to loss prevention, allowing them to make informed decisions on how and when to intervene in cases that, left alone, might be of sufficient complexity to produce significant losses.

 

Loss Control Benefits

Beyond helping reduce litigation, using AI to analyze workers’ compensation claims can also give employers a better understanding of their workers’ compensation risk, information they can use to reduce exposures over time.

An application with built-in “explainability” — features allowing employers to understand the claim attributes and characteristics that the models identify as contributing to heightened risks — provides added value.

Those factors might include the injured employee’s occupation, wage level, age or geography. “Explainability” features will also clarify to the employer how much weight the AI models assign to the different characteristics, as well as the claim’s potential risk for litigation.

Armed with information about factors influencing their litigation risk, employers can start risk-mitigation efforts earlier, perhaps taking a different approach to managing individuals that exhibit higher risk characteristics.

As businesses run the models over time, gaining more and more information about the attributes most common to higher litigation risk, they can develop strategies and processes to help mitigate litigation risks associated with particular types of claims or employee profiles.

 

Measuring the Performance

To truly understand the benefit the AI application is delivering, performance must be tracked across multiple perspectives. Of course, information on the number of cases the technology determined to be high risk and the number of cases closed successfully without litigation are important leading indicators.

It’s also necessary, though, to examine lagging indicators, such as how the current volume of litigation compares with tomorrow’s — or how yesterday’s volume compares to today’s. One goal, of course, is to constantly reduce the volume of litigation. But it’s essential to connect that metric to total cost and to identify whether the total cost of the workers’ compensation program is decreasing on a normalized basis.

Many of the specific metrics to be tracked will depend on the employer’s specific objectives. Some companies, for example, might have particular objectives about reducing litigation or litigation expenses within certain operating entities. Ultimately, though, the metrics used should support the effort to better manage workers’ compensation claims to reduce total cost of risk.

 

Technology and Competitive Advantage

As in other areas of business, using technology to address workers’ compensation claims and other risk areas to reduce cost of risk can be a source of competitive advantage.

Businesses that leverage the power of AI in their insurance and risk-management programs can help better control their fixed insurance costs as well as their variable retained loss costs. The applications will only get better over time, increasing the potential benefits.


[1] Six Ways to Reduce Workers’ Comp Claims Using AI, Clara Analytics