The State of Artificial Intelligence and Workers’ Compensation Claims

July 28, 2020 – Michael Marsh

An invitation to attend a webinar was received recently, sent by a company that specializes in Artificial Intelligence software directed at the workers’ compensation claims industry. Software that can improve claims services, recovering worker healing times and quality, and improve efficiency have come a long ways in the last couple of decades. However, many of the ‘offerings’ initially (during the demo and sales process) show benefits to the organizations licensing the software. For a variety of reasons we’ve seen these fail time and again when implemented. Sadly, some of the ‘offerings’ have proven to be counterproductive…and quite expensive.

Thinking back, the year was 1991. As the Home Office Vice President of Claims for a SoCA Workers’ Compensation and GL insurer, with nearly a billion dollars of run off claims exposure (acquisitions and roll ups) in 38 states, I was asked to participate in the company’s computer system reengineering efforts. Might have had something to do with the fact that I had some programming experience from college, and I was about half the age of the remainder of the executive staff? 

For those reading this below the age of 40, you may not believe it, but this was before the personal computer invaded the insurance office. We were an IBM mainframe shop. In the process of ‘reengineering’, I’ll never forget. We had two of these new IBM machines in Accounting. Everyone was intrigued with the ability to create this thing called a spreadsheet and to have a printer connected, right in their office.You could make something on the screen, spreadsheet or in this new thing called a word processor…and print it right at your desk. We dreamed of the miles to be saved of walking to and from the shared main frame printer. Ah, but I digress.

The reengineering included consideration of a fairly new product from IBM called Image Plus. We of course had never seen it in action, there were only a couple of installed clients in the country. One of those installations was at USAA in San Antonio. TX. At the invitation of IBM, we were taken to San Antonio and given the full tour. There we saw what to many was amazing, if not otherworldly technology throughout the huge campus. We were told that the days of paper were over. We saw a demonstration of the scanning, indexing and internal routing of documents in action. By 10:00 am, every piece of incoming paper mail had been processed and sent to the shredder. Except for legal documents, notice of Mediation and other bulky documents such as 500 page legal papers with attachments. But we heard over and over, in the next few years there would be NO PAPER in the insurance industry.

That was nearly 30 years ago. Is the insurance and claims industry three decades later in the NO PAPER environment? Certainly not. The sales pitches and promises made simply didn’t have the “in the trenches” experience to know that technology was not yet ready to rescue the industry from its “wasteful, anti-environmental” use of paper. Has there been significant improvements in efficiency and reduced reliance upon paper documents? Absolutely. But the reality was far less nirvana than the sales brochure and presentation.

I learned an important lesson through this process. Technology “promises” need to be met with pessimistic optimism. Hope for good results while expecting reality to be delivered far less than promised by the shiny shoes person. The contract and terms of service are very important. When the excrement hits the fan in your shop once users learn that the “promises” were not realistic, many times the shiny shoes are gone and calls, e-mails and other paths lead only to customer support.

Which leads me to the announcement of Artificial Intelligence (“AI”) as the ‘new tool’ to help the workers’ compensation industry. The sessions learning goals, fixing and optimizing the claims process:

The expert panel explains how AI can: 

  1. help you identify the right provider for a case and steer the injured worker to that provider
  2. help busy adjusters easily spot potentially troublesome cases and manage them better, from start to finish
  3. continuously optimize your network of providers, so you can be sure to have the right provider working with the right worker at the right time.

AI has been around a very long time. A brief article about AI and machine learning can be read at . Humans have been curious about, some obsessed with, finding a way to make the machine think like a human. This has been glorified and almost made to seem real in the media and movies. Recall the Terminator franchise or 2001 – A Space Odyssey? The assumption that has been in the narrative for decades now is that machines will learn enough from human programming to become self-aware…and we should as a human race be afraid, very afraid.

I’m less concerned about the near term death of all humans at the hand of awake machines than some others. Machine learning (or AI) is essentially a series of algorithms that get input and then execute a result just as they are told to do by their programmers. My lessened fear of machine learning / AI is not so much about the machines…except for some misbehaving apps from Washington, (recall the blue screen of death in the XP days?) the machines do what they are told to do, when they are told to do so, by their “masters”. My confidence of the survival of the human race with respect to AI is rooted in the inability of most humans to think/evaluate/program the machines to think in the sophisticated way that the human brain works. Humans are generally not Yes/No switches. We understand the gray areas, innuendo, rumor. Machines don’t and at this point cannot. Machines are simply Yes or No. Creating algorithms sufficiently sophisticated to tie together thousands, if not millions, of Yes and No responses is simply beyond our ability to understand and program at this point in human history.

Around 1992, when we began using a personal computer seriously, using this new miracle software (spreadsheet) and data from the mainframe, we were able to begin building a predictive model of claims costs. Without a huge scope document sent to those strange men and women in that locked room way back there we could begin studying our numbers. Now. Not in 6 – 9 months waiting for a new report to come out of COBOL-land. Now. People learning about the business…using machines.

It was a positive time, optimism about how computers would help revolutionize the insurance claims industry. There was always talk of how the predictive modeling was going to obviate the need to do case reserves on claims. We heard, if we have the type of injury, the severity, the treatment codes, we’ll know with xxx% (insert your own high number, cause they were all wrong) certainty what the book of business will cost by state. Just a few reports from the data in the system and we’d know what next year’s rates would be…no longer dependent upon those distasteful people back in the Claims Department. Underwriting, Loss Control, Compliance, Accounting…seemed like everyone share the vision of a group of automatons in the claims area…the computer will let us hire less of those people and we’ll have better, more timely information and none of those last minute, pre-trial reserve evaluations. The AI machines will guide us.

They were wrong.

It has been a couple of decades of promises, optimism and hope for Artificial Intelligence to rescue the industry. Have there been some moderate efficiency improvements? Of course. Most insurance companies and TPAs I speak with echo that sentiment. But let’s be honest. There hasn’t been that much investment in AI. Much of the industry has been trying to figure out how to more effectively and effectively pass claims and overhead fixed costs on to the customer in the ever-increasing rates. 

For at least 10 years companies have been moving to the work-from-home claims model. This has picked up steam and much press coverage with the COVID crisis of 2020. Is the work-from-home motivation to achieve better service levels, develop closer and productive relationships between the adjuster/examiner and the recovering worker and the treating legitimately injured workers with respect and dignity? My take on it for years has been that work-from-home is a desire to simply reduce fixed costs. Fortunately in the many years that have passed, imaging has improved and the costs have come down sufficiently that paper can actually and efficiently be changed to digital images to allow people to work at home in their jammies. What happens when the indexing is incorrect, images lost, misdirected or damaged? That topic is for another day.

So is ‘off the shelf’ AI software now sufficiently sophisticated to?

  1. help you identify the right provider for a case and steer the injured worker to that provider
  2. help busy adjusters easily spot potentially troublesome cases and manage them better, from start to finish
  3. continuously optimize your network of providers, so you can be sure to have the right provider working with the right worker at the right time.

After many claims audits, appearing in extra-contractual litigation as a claims process expert and viewing tens of thousands of workers’ compensation claims, my experience has been that the best predictor of a claim resolution is an experienced adjuster/examiner that:

  1. is empowered to make claims and financial decisions, 
  2. knows their jurisdiction well,
  3. has strong administrative and systems support, and 
  4. has a low ongoing case load.

Takeaway: Buyer Beware

Sign Up

Sign up for more great content delivered directly to your inbox.