Grow property & casualty revenue without adding expense
Get work done faster, from underwriting to claims processing
Free up employee time for higher-value work
Codify and streamline property & casualty processes
With stalwarts and insurtech startups alike
Improve customer response time
For years P&C insurers have been trying to automate various processes with limited success. That’s because these attempts often used rule- or template-based approaches that don’t work well with the unstructured data that dominates P&C insurance processes.
Consider the P&C underwriting process. Underwriting involves collecting various documents describing the property in question for a property insurance policy, so the underwriter can accurately assign a value to it and calculate replacement costs.
Customers may submit various forms to aid in this value analysis, many of them in unstructured and sometimes complex formats, including PDFs and Word documents as well as photographs. Typically, a front-desk team would review these documents as they arrive, typically by email. The team would read each document, looking for relevant information, then rekeyed into a downstream system that collects the data for the underwriter.
The data entry job is labor-intensive, time-consuming, and monotonous, making it prone to error. That’s why P&C providers have been trying to automate it, initially by using rule- and template-based approaches that look for specific keywords.
Such an approach is all but futile given all the unstructured data involved in the P&C underwriting process. Templates and rules work only on highly structured data; they rely on the data being in the same place from one document to the next.
Optical character recognition (OCR) is another approach often touted as a P&C insurance process automation solution. OCR is a machine learning technology that can be used to convert documents such as PDFs into a machine-readable format. While that’s useful, it doesn’t address how to extract the relevant data.
Using robotic process automation (RPA) for insurance likewise suffers from limitations when it comes to unstructured data. As its name implies, RPA uses software robots to repeatedly perform highly structured and repetitive tasks involving the same keystrokes.
But that’s not how the P&C underwriting process works. It requires a human to read documents and make judgments about which data to extract. Any other P&C insurance process involving unstructured data, which is most of them, will suffer the same fate regarding OCR and RPA. (RPA can complement intelligent document processing in insurance automation, more on that below).
Indico’s approach to intelligent document processing, Intelligent Process Automation (IPA), is fundamentally different from RPA and templated approaches because IPA can understand document context much like a human does. Our IPA platform is built on top of a database containing 500 million labeled data points – enough to enable it to understand human language and context. It would take the largest P&C insurance carrier years to accumulate that much data and build its own model.
We apply AI technology known as transfer learning enabling users to create custom models that can automate virtually any P&C insurance process. It takes only about 200 documents to train a process automation model that delivers around 95% accuracy. Another key point: you don’t need data scientists to build AI process automation models. Rather, it’s the business people involved in each process who train the automation models – those who know the processes best. (For more on this, check out our Intelligent Process Automation page.)
Automating the P&C insurance underwriting process requires reviewing numerous documents, with relevant data extracted and entered into a downstream processing system. An effective intelligent document processing solution automates processes by “reading” these documents much like a human would find the relevant data. It can also automate data extraction and data entry, saving an untold number of hours to dedicate to more valuable work.
Numerous processes may be involved in servicing a P&C insurance policy over its lifetime, including:
Many of these processes involve dealing with various forms of documentation, making them ripe for intelligent document processing.
The claims process is rife with documents coming in from various stakeholders, including customers, adjusters, brokers, appraisers and more. The documents may arrive via email, fax, websites, or traditional mail. Here again, the traditional manual process requires claims representatives to review the documents to find relevant data and manually enter it into the claims processing system.
The P&C claims process involves numerous steps including:
Many of these steps are transactional in nature, involving discrete steps and document reviews. An intelligent document processing tool can automate many of the steps in the P&C FNOL process including the document review process, extraction of key data points, and setting up the claim in a P&C claims management tool. It can also validate that all required data is present before sending the claim to an adjuster. Simple claims may be automated end-to-end based on pre-defined business rules, perhaps with only a final review and sign-off required at the end of the process.
The claim adjudication and subrogation process likewise consists of a number of steps that can be at least partially automated, including:
Applying intelligent processing to the numerous documents involved can streamline these processes and speed resolution of the claim, thus improving customer satisfaction.
For some P&C insurance automation use cases it may make sense to use robotic process automation to complement IPA.
RPA works well on processes that are highly deterministic in nature and involve structured data. In that sense, it’s well-suited to automating repetitive tasks, making a process less labor-intensive for humans. IPA, on the other hand, is able to automate processes that involve unstructured data.
A common IPA and RPA use case, then, is to use IPA to “read” unstructured data and translate it into a structured format before handling it off to an RPA tool. For example, the RPA tool may perform the initial document intake, then send the documents to an IPA tool for classification and data extraction. The IPA platform can then translate the extracted data into a structured format, such as a spreadsheet. The RPA tool can then take the now-structured data and automate the process of entering it into a downstream system, such as a claims processing system.
It’s challenging for P&C providers to keep up with business requirements under the best of circumstances. But it’s imperative if firms are to achieve digital transformation. Indico’s intelligent document processing platform helps you take a big step in the digital transformation journey while delivering significant benefits, including:
Automation enables P&C adjusters, appraisers, examiners, investigators and other employees to be more productive, enabling the company to increase revenue without adding headcount.
Automating P&C insurance processes empowers companies to get work done faster, even while increasing accuracy.
By automating mundane tasks, you can free up employee time for more rewarding work that’s also more valuable for the company.
Part of the value of a P&C insurance process automation exercise is codifying processes that may have existed for years with no formal agreement on how they are supposed to work. It’s also an opportunity to streamline processes to make them more effective.
Intelligent process automation ultimately makes your organization more competitive, putting you on equal footing with the most nimble insuretech startups and largest industry players alike.
Customer expectations are at an all-time high. Intelligent automation allows P&C providers to exceed client demands by improving the speed and accuracy they can react to customer needs.