Automating the direct bill receivable process for a global insurance agency

Case study

Automating the direct bill receivable process for a global insurance agency

Insurance agencies have multiple ways to bill their customers. On one hand, they accept policy premium along with their commission from the insured. On the other, the insured makes premium payment and commission payment directly to the insurance carrier, which in turn pays the latter to the agency through a commission statement. This second way of billing is termed direct billing.

As a part of direct billing, generally the finance team is responsible for maintaining a record of estimated income from commission, receiving the actual income from the carrier, tallying both to ensure the estimated income and actual income matches, and entering the commission into the agency management system (AMS).

Manual nature of this pretty straightforward process leads to the following setbacks:

  • Extra human resources are engaged to perform high volume repetitive tasks.
  • Cost to the company increases.
  • There is an additional risk of manual errors while reconciling commissions earned.

Robotic Process Automation (RPA) is the best fit for automating such repetitive tasks. Sample these:

  • The cost of a robot can be as low as a ninth of a full-time equivalent (FTE). 
  • Robots work as much as 10 times faster than their human counterparts.
  • Software robots are available 24×7, perform repetitive tasks rapidly, more efficiently, and with zero or low errors.

Requirements

Our client — a leading global firm engaged in providing insurance brokerage and consulting services with a market cap of over $10 billion — used to employ manual resources in their direct billing process. The need was to automate it to save costs, reduce time and error, and efficiently reallocate human resources to more vital jobs.

The client reached out to us to with specific goals to:

  • Reduce the FTEs in the process so they may better utilize them
  • Ensure the automation does not violate the turnaround time (TAT) metrics of the process
  • Ensure the automation does not enter inaccurate financial data into the AMS

Challenges

Our client received commission statements from ~65 carriers in a month. Now, multiple carriers came with multiple commission statements in the forms of Excels and PDFs. Extracting data from Excels are pretty simple and straightforward. But the varied document structures of PDF commission statements from multiple carriers made it a task for the extraction technology to extract financial data accurately.

Solution

We suggested building an RPA solution to support our client’s direct billing process by automating extraction of financial data from commission statements and entering it into the client’s AMS. The RPA solution incorporated would reduce the manual effort required in the current process.

The following steps were automated:

  • Extraction of data from the carriers’ commission statements
  • Entering accurately extracted and verified financial data into the client’s AMS
  • Recording technical and business exceptions the bot encounters in the automation process

Tech stack

How Our Solution Helped

Our solution automated a critical billing process which saved time and money for the client as well as increased process efficiency.

Overall Approach

We identified that the key step in the automation journey would be data extraction from the insurance carrier’s commission statement. If the statement is in a simple Excel form, the bot extracts the data, verifies it and enters it into the client’s AMS.

If the commission statement is in a comparatively complex PDF form, to address the challenge of variations in the document structure of the PDF commission statements, rules-based document extraction templates are created for each commission statement (E.g: extraction based on a specific data position). The ABBYY FineReader extracts and classifies the data it is sure of, from the templated statement, and sends it to the bot for entering it into the AMS.

There could be a 1-2% chance of inaccuracies in an Optical Character Recognition (OCR) technology. If the ABBYY FineReader is not sure of some extracted data (in terms of exceptions, anomalies, discrepancies), it sends it to the Verification Station that tallies the extracted data with the original and highlights the inaccuracies. A human reviewer handles the highlighted data and sends it to the bot after proper verification, post-which the bot enters the verified data into the AMS.

If any business or technical exceptions, anomalies or discrepancies are further encountered while performing the process (whether with Excel or PDF statements), the bot logs comments in an exception tracker.

The following diagram presents an architectural overview of the RPA solution implemented and how it works:

Here is how the process flow looks like:

Results

  • The RPA implementation reduced the number of FTEs by ~50%
  • Approximately 90% of the entire direct billing process was automated
  • There was no negative impact on the process TAT metrics 
  • There was a significant reduction in human data entry errors

Talk to us