SortSpoke Blog » Latest Articles
blog featured image
Jan 23, 2024 SortSpoke

AI Document Processing Helps a Fortune 500 Insurance Co. Leverage it's Unstructured Data

This anonymous case study is part of an ongoing series that gives readers a behind-the-scenes look at some of the most challenging, impactful, and sensitive business problems in the insurance industry solved by SortSpoke. We keep the clients and their companies anonymous so they can share full details.

Managing high volumes of incoming documents is a continuous challenge for insurance companies. It is necessary for many different workflows, including during the claims process and especially during underwriting. Managing the intake of high volumes of data is something every sector struggles with, but what makes the insurance industry different is the greater variety of unstructured data it must handle.

What is unstructured data?

Unstructured data is any digital information lacking a formal structure—meaning it is not easily enterable in a database or spreadsheet. In insurance underwriting, that frequently means text in documents (like PDF files), emails, loss runs, statements of value, financial statements, or other scanned paper documents, including handwritten ones.

Historically, data in all of these formats had to be processed manually by human staffers. That is obviously very time-consuming and inefficient. And transcribing is very prone to human error. Those factors can seriously affect the quality of underwriting decisions, and, therefore, corporate profit margins and strategy.

For one Fortune 500 insurance company, the headaches surrounding manual document processing had grown so large that they could no longer be ignored. This case study explores how they came to identify the nature of their document intake challenges during underwriting and how they addressed them with the help of SortSpoke.


Decentralized corporate structure made standardizing workflows difficult

The assistant vice president for analytics within a major division of this Fortune 500 insurance company was concerned with the problem of manual document processing and its impact on the entire company’s operations. While her division focused primarily on casualty insurance for enterprise clients, this AVP’s primary responsibility was uncovering organization-wide problems and identifying digital solutions for them.

This was a lucrative company but a very decentralized one. It contained 36 different divisions. Some were spun off from the original company, and others were acquired. Each insurance division had a degree of autonomy in how they operated.

“We'd like everybody to get along, but they don't have to,” the AVP said. “They have their own leader, their own IT team, and sometimes their own strategic initiatives. If you're a smaller division and you need to pair with three other divisions, they'll do stuff across business units that way if they don't have quite as many resources.”

Over time, that decentralized environment inevitably led to divergent business processes, especially around document intake, processing new accounts, and the underwriting process—anything involving information technology. “So when we’ve acquired companies, they usually have a whole [IT infrastructure] stack because they were an individual company,” the AVP said. “So they get to keep their IT. And that was what happened across all the different units, and why some of us have IT, and others don't.” That was where her role came in: identifying solutions where resources were limited.


Effecting change was not easy

Generally, her team would look at what IT systems and associated business processes were about to be retired. They looked to see what could be modernized and improved. Document processing leaped out at them.

Their greatest concern was the excessive manual labor needed to process files and scanned documents and wrangle the unstructured data they contained into something their systems could work with. The problem was that only some higher-ups agreed with her on priorities in their decentralized business.

The company’s bottom line was solid. They were not in cost-cutting mode. It was not necessarily everyone else's top priority because the AVP highlighted a pain point in their underwriting process that could be improved. She needed to figure out a way to make clear how modern, AI-driven document management could be a strategic initiative.


Looking for solutions

“I was looking specifically for data extraction vendors, but I talked to almost every applicable insurtech in the industry,” the AVP said. She wanted to know everything about “what's out there and how we could possibly apply it.”

Her team came across SortSpoke and realized they could use it to enhance workflows throughout the underwriting process. With reliable unstructured data collection and processing, her team might be able to develop “an enhanced user experience” for her business unit and others within the company. With SortSpoke, her insurance raters “can easily make corrections and incorporate the data without having to send it back and say, ‘This is wrong,’ and wait for it to be corrected.”

The versatility of SortSpoke’s data extraction capabilities made it the best choice for their needs. But now, this AVP had to get the buy-in from her wider organization.

For a deeper dive on this topic, get our free Evaluation Checklist with an 18-step decision framework.


Emphasizing the impact of underwriting errors on major casualty claims

A well-run casualty insurance business will have plenty of checks and balances to ensure they are not carrying too much risk for claims of that scale. Their business remains profitable if those claims are high-risk but low-frequency. However, their risk assessment is only as good as the quality of the data collected and analyzed during underwriting.

If new Federal regulations or precedent-setting legal suits go through, there could suddenly be an emerging trend for a whole category of related lawsuits against this company’s clients. The AVP realized this was the compelling business case she needed for automating their document processing. “Those emerging trends, those lawsuits that seem to go across a bunch of different states, those are the ones that create urgency for us. Our leadership wants to know, ‘What's going on? How are we exposed? Where are we exposed?'” Those questions could be answered quickly and easily with higher data quality at intake.


Problems working with other AI document processors

The AVP thought that SortSpoke and its ability to process unstructured data had a unique benefit in capturing valuable information from scanned documents in the underwriting process. “Our raters will get applications that don’t always fit into a neat box. There might be multiple layers of documents all piled together. There might be overlapping property or general liability insurance layers because brokers don't want to fill out multiple applications.”

However, she soon discovered that their liability insurance division was also investigating AI document processing. Corporate IT wanted the casualty division first to try one of SortSpoke’s competitors. Unfortunately, it did not fit their needs. “It was great for very structured documents. But if anything was off by a millimeter, you have to create a whole new template. It's not terrible. It works well for high-volume, identical documents. But not for what I was trying to do.”

So, she got the green light to go with SortSpoke.


Deploying SortSpoke across the organization

To prepare for SortSpoke’s deployment, the AVP developed more than a dozen use cases with stakeholders from different divisions throughout the company. This insurance company's culture of digital transformation was finely tuned. They were not just going to throw new technologies into the organization, cross their fingers, and hope it was adopted. “We need to integrate everything into an overall workflow. We don't just try to give somebody a tool to just go about their merry way.”

The deployment was not all smooth sailing, though. The AVP had to deal with constraints on the IT side. Because they were using SortSpoke as the engine powering bespoke software solutions for different business units, they needed a good deal of internal development and design work done. 

However, with assistance from SortSpoke engineers, she quickly made the necessary modifications for the tools to be put in place throughout the organization. She appreciated that hands-on approach. “You guys are my favorite people. SortSpoke is my favorite vendor by far,” she said. “I don't say that lightly. I work with a lot of vendors.”


Rapid adoption and clear benefits

Most staff were impressed with the SortSpoke-powered tools right away. The deployment was so successful that other business units within the insurance company soon asked the VP for assistance deploying it for some of the other use cases she identified when initially scoping her document processing project.

“I've been using SortSpoke for many different kinds of analytical stuff. We have two other projects where we want to incorporate SortSpoke with both submission intake and claims intake.”  What she originally sought for its capabilities supporting underwriting application processing had other clear uses for this insurance company.

While some workers were initially resistant to learning the new tools, adoption grew quickly after users saw how easy it was to feed SortSpoke documents for processing and analysis. “We've been able to use SortSpoke for a variety of different types of work. Both short-term for analytics or looking at long-term integrated approaches, like extracting data specifically for analytics purposes.”

This Fortune 500 insurance company is now seeing steady improvement in its underwriting, claims, and other processes. Their metrics show their intake process is four to five times faster per underwriting application. That has resulted in a 40 percent increase in written premiums.

“I really like SortSpoke because of the ability to extract information from unstructured documents in a very user-friendly, interactive way,” the AVP said. “No other product out in the market does that. Others want you to create models, and they want to be able to pull information out, but they don't have the same kind of interface that allows you to correct through workflow, which SortSpoke does.”


See how SortSpoke can improve your underwriting process

Book a risk-free, no-obligation discovery call with our team today.