Process Mining Can Help Reduce Process Friction in the Insurance Claims Process

process mining can help reduce process friction in the insurance claims processOptimizing the claims process is critical for controlling operating costs, increasing working capital, and improving customer satisfaction. But many insurance companies struggle to make significant improvements because they:

  • Can’t clearly identify the execution gaps and root causes of the “friction” in their processes 
  • Can’t effectively implement and monitor the actions that close each gap 

This post briefly shows how you can use the Doculabs Claims Execution Management Solution to identify friction in the claims process and remove or reduce it. You can see a demo here. I’m going to focus on the initial phase, discovery, of a process optimization project.

Our clients typically first use the application to discover the process friction points and the root causes of those inefficiencies. Typical friction points and their causes include assignment to the wrong adjusters, lack of process standardization, rework, and lack of automation. By knowing the root causes, you can make changes to the process to resolve the issues. And then after making the first round of enhancements, they monitor and continually improve their claims process. 

Our claims clients typically want to: 

  • Improve customer satisfaction 
  • Reduce claims payout 
  • Improve operational efficiency and operating costs 

I’ll focus on customer satisfaction and excess payout in this post. 

Introducing the Doculabs Claims Execution Management Solution 

The application uses the Celonis process mining platform to provide an end-to-end view of the claims process by extracting data from a client's source systems, such as Guidewire or Duck Creek. We can also extract event log data from document capture systems such as Kofax and automation platforms such as Appian and Pega. 

Where should you start process discovery? Start by looking at the big performance indicators at the top of the application – here we have claims count, payout, duration (a proxy for efficiency), and satisfaction (in terms of Net Promoter Score). In this case, we currently have 7,400 claims with a $41 million payout, an average cycle time of 71 days, and an average Net Promoter Score of 47, a good score. We’ll be drilling into satisfaction and payout, to determine where we have low performance, what causes it, and what we can do about it. 

Investigating the Claims Process and Its Variants 

The next step is to take a look at the actual claims process. In the center of the new application window below you can see the most common path in the claims process.  

Graphical user interface, application, email
Description automatically generated

While this is the most common path, there are likely many variants or divergences from that path. The image below shows all the variants that make the whole process look like a spaghetti bowl. The variants are where you start digging to find and remove process friction. You’ll find rework, non-standardization, lots of delays, reassignments, and many other inefficiencies and their causes. This allows our clients to save hundreds of hours diagnosing broken processes.  

Example #1: Investigating Rework, Cycle Time, and Customer Satisfaction 

Let’s drill into an example causal chain that runs from claims process inefficiencies and their root causes to the resulting poor business outcomes that can be improved. We’ll look at the impact that rework has on extending a claim’s duration, and then the impact that those laggard claims have on customer satisfaction. The rework might be needed because of supplemental damages, maybe engine damage in addition to body damage from a car accident, that are uncovered after the initial inspection. This requires that the process to loop back to earlier steps and perform steps already completed or new steps in the process.  

In the example below we focus on the claims that require rework and that took 95 days to close. We see that customer satisfaction as measured by NPS (Net Promoter Score) drops from 47 to 27, well below the auto insurance industry benchmark of 41. 

 

Example #2: Investigating Adjustor Misassignment and High Loss Adjustment Expenses 

Let’s now look at one of the most common problems we find in the claims process. Assigning claims appropriately to the right adjustor and skill level is critical. If complex claims are misassigned to inexperienced adjustors or those without the right areas of expertise, several kinds of problems arise, causing customer dissatisfaction, increased operating costs, and increased payouts. In our example analysis, we drill down to determine how many claims were misassigned to inexperienced adjustors and view the impact it has on claims payout and loss adjustment expenses.  

In the Application example below, look at the two red circles. We are focusing on 1,684 complex and highly complex claims that were assigned to “Level 2” adjustors, those adjustors with some but not a lot of experience. The result is that the excess payout has increased to over $900K. (The excess is defined as the difference between the total expected payout of $8.7M and the total actual payment of $9.6M.) This kind of analysis allows us to identify the situations that lead to higher-than-expected loss adjustment expenses early in the process, e.g., at the adjustor assignment step. 

Summary of Problems and Solutions 

In our example scenario with the Doculabs Claims Execution Management Solution we found multiple variants of the process, significant rework and delays causing customer dissatisfaction, and misassignment of adjustors causing high loss adjustment expenses. If we continued to drill into the scenario with the Solution, we’d find more specific evidence for the following execution gaps: 

  • Claims are reassigned or reworked because of incomplete or incorrect data and document collection 
  • Intake assessments are inconsistent with the corporate guidelines and standards and inadequate or wrong information is initially collected 
  • Assignments to adjusters and shops don’t match the skill level of the adjustor or the appropriateness of the shop because of lack of relevant adjustor and shop data 
  • Many activities are manual and paper-based rather than automated and digital 

The recommended improvements are the following: 

  • Implement standardized AI-assisted intake and assignment, including intelligent document and data capture, guided e-forms, and workflow that uses updated information about adjusters and shops to assist with assignments 
  • Apply automation and digitization to other specific manual, paper-based steps in the claims process 
  • Provide robust procedures, guidelines, and training to ensure that all offices conform to company guidelines and standards during intake and other claims processing activities 

The next step is to use the solution as a guide to help Implement and automate the recommended improvements. Then, after implementing improvements, you can continue to monitor overall performance along with the root causes and implemented improvements, both real-time and over longer periods. When necessary, the Solution can also make real-time process improvements using machine learning and automation functionality. 

Conclusion 

The above is a simple example but more generally the Claims Management Execution Solution will:

  • Reduce inefficiencies caused by staffing and resource issues, including allocation and training 
  • Reduce claims overpayment and loss adjustment expense. 
  • Reduce long cycle times and quality issues due to lack of standardization, digitization, and automation.  
  • Reduce rework, redundant processing, and reopened claims. 
  • Improve customer satisfaction, retention, and upsell 

Ready to improve your claims processes? Call us at (312) 433-7793 or fill in the contact form below and we'll be in touch ASAP. 

Rich Medina
Rich Medina
I’m a Principal Consultant and co-founder of Doculabs, and the resident expert in using ECM for information lifecycle management.