patient direct pharmacy logo
August 8, 2019

How AI Advances ‘No-Touch’ Resolution of Workers Compensation Claims: Opinion


Many human hours are needed to sort, analyze and process workers compensation claims—the vast majority of which are low in value.

The U.S. workers comp insurance market is forecast to earn revenues of $55 billion in 2019. To be competitive, it’s important for a carrier to be as efficient as possible with claims. Fortunately, technology in the form of artificial intelligence (AI) is now available to help insurers with their lower-value workers comp claims.

Looking at the claims situation in a bit more detail, a study published by the National Council on Compensation Insurance (NCCI), published on Sept. 25, 2018, found that the biggest category for workers comp claims in 2016 was for those of less than $5,000, accounting for 27 percent of claims.This was followed by claims of $10,000-$25,000 (22 percent); $25,000-$50,000 (18 percent); $5,000-$10,000 (14 percent); $50,000-$100,000 (13 percent); $100,000-$500,000 (7 percent); and $500,000 and above (less than 1 percent).

Rupin Mago

In the segment of low severity costs, our experience is that even though it is hardly worth it for an insurance company to invest much time analyzing these claims, there are vast armies of experts involved in multiple touch points in the workers comp process, carrying out low-value and highly repetitive manual tasks due to a lack of digitization.

A typical European insurance company that operates within a single country will have 30 to 50 staff working on sorting the mail into the right category—claim, medical report, medical bill, etc.—and then scanning in information. An additional 150 to 200 people will be tasked with extracting information such as invoice costs, claims numbers and doctors’ reports and putting them into a digital format. When it comes to resolving claims, there can be multiple "to-ing and fro-ing” over a lengthy period of time—from doctors’ reports and authorization of treatment through to paying the final bill, which can involve around 800 people.

In our experience, U.S. insurance companies are far bigger and the uptake of software tends to be lower, so the process is even more labor-intensive. As a result, we are finding that there typically are twice as many people involved due to claim quantity.

Clearly, the system is currently highly labor-intensive and there are far too many people employed in the industry who carry out administrative functions. AI is one of a number of evolving technologies that have the potential to revolutionize the way the insurance industry processes claims.

Transforming Claims Processes

When it comes to claims, there are a number of ways in which AI transforms the administrative back-office processes. For example, machine learning also analyzes the vast amounts of text-based communication that is omnipresent in claims management.

Dependent on the format of an invoice, the invoice details could be in different locations in the text. AI reads and understands the text at superhuman speed and identifies the vital information such as invoice details, as well as other data like claims number and the person’s date of birth and address.

After absorbing and analyzing this raw data, AI can be used to scan other data sources to resolve a claim.

AI-fueled claims management provides recommendations on a next-step basis if straight-through processing is not possible in complex cases. Currently, expert examiners will mostly use their own experiences to decide what to do next. A person with three years’ experience versus someone with 15 years’ experience may choose different next steps. AI can help expert examiners access data from a variety of similar cases and provide a recommended next step based on that historic data, so the carrier achieves more consistency in the processing of claims.

The software can find out information such as whether a person is insured and whether the person is insured for a particular peril. Using this technology reduces the possibility for human error that can occur when there is manual inputting, making claim handling far quicker and also more accurate.

In claim cases where manual intervention is normally required, we are finding that it takes 12 to 30 minutes to resolve the claim with AI technology. Insurers typically cluster claims by complexity, and initial tests suggest that AI can help bring about vast reductions in turnaround times due to "no touch” resolution of claims with low complexity and enable straight-through processing of up to 60 percent of those cases.

Due to a lack of necessary technological capabilities, many insurance companies are currently unable to implement automation in a useful way—i.e. at various points (routing, triage) along the claims cycle. "The challenges include a continued reliance on slow and unwieldy systems and the difficulties of moving old—sometimes very old—books of business over to new platforms,” said a PwC report published in March 2018. The potential time savings insurers could achieve are clearly considerable.

Insurance companies are the custodians of data goldmines, and they can use AI for a range of functions beyond claims to analyze this complex information and create highly valuable advanced business insights for insurers.

A 2019 NCCI survey of insurance executives found the top concerns keeping executives awake at night were: the changing workplace/workforce; profitability/premium; medical costs; political/legal environment; and the future of workers comp.

AI has the potential to address a number of these issues.

New technology and ways of doing business mean the workplace is changing rapidly, as is the nature of employment, so insurers need to have a holistic understanding of this rapidly changing environment. They can use AI and machine learning to aggregate claims information, which allows them to identify trends such as when claims are likely to be submitted by sector, the types of injuries and average recovery costs for particular injuries.

The technology allows the insurer to factor in a claimant’s medical history. For example, has the claimant had heart problems, and do the records show if they are a slow or a quick healer? By integrating this information, rather than just looking at the average time it takes a person to recover from a particular injury, the insurer can underwrite in a far more accurate, personalized and ultimately profitable way.

Freed from the more mundane inputting tasks, insurance professionals can concentrate on higher-value customer service that puts a strong emphasis on empathy. For example, one service provider allows insurers to contact claimants every week throughout the claims, treatment and recovery process and rate their experience on a scale of 1 to 10. AI can track the customer responses each week, and if the sentiment is starting to go negative, it will flag this to the insurance company, triggering a human intervention to ensure the customer gets back on track to a more satisfactory experience. Insurers that deliver high levels of service are much less likely to lose customers through churn.

Advanced data processing empowers insurers to have a much more comprehensive overview of claims and their customers’ risk profile, which can identify outliers that may be a sign of fraud. The use of very segmented information allows for the creation of bespoke policies and prevention programs for customer segments.

Demonstrating Value

Looking to the future of workers comp, one major concern is a trend for large companies to self-insure their employees for workers comp and outsource the claims processing to TPAs. If insurers want to pre-empt or dissuade large companies from self-insuring their workers comp programs, they need to be able to demonstrate that they can deliver a far quicker, cheaper, more efficient and personalized service—and that is where technology and automation can help.

We are discovering that uptake of software to automate workflows in the workers comp space is still low, yet we are confident the use of technology will be one of the key factors to decide which insurers flourish and which fall by the wayside in the years ahead. As a McKinsey report published in May 2018 found, by 2030, "claims for personal lines and small business insurance [will be] largely automated, enabling carriers to achieve straight-through-processing rates of more than 90 percent and dramatically reducing claims processing times from days to hours or minutes.”

map-markerhistorycrossmenuchevron-down linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram