The Situation

A major regional Property & Casualty insurer with a large portfolio of commercial property policies was challenged by a heavy volume of underwriting and claims documents, exceeding 400,000 documents annually in a “normal” year. During an intensely active natural catastrophe (CAT) season, document volumes could soar into the millions, creating an operational “perfect storm.” The carrier’s dedicated teams, the backbone of their service, were stretched thin, managing a flood of faxes, mail, and emails, all processed by hand.

The Challenge

Relying on manual processes had created a severe bottleneck, with both claims indexing and commercial underwriting operations straining to keep up with time-consuming manual tasks. This resulted in inefficiencies, inconsistent data, and a reduced ability for their professional claims administrators, adjusters, and underwriting staff to focus on complex, high-value work. These teams had ambitious growth goals—to increase capacity 70-80% in claims and 60-80% in underwriting—setting targets that seemed unattainable with their current workflows and processes.

A major impediment to success in meeting objectives was the lack of standardization in document indexing. With a staggering 300+ different document types in use across the company, with disparities between documents handled by field adjusters and internal admin teams, maintaining accuracy and efficiency was a daily challenge. 

Key documents that are vital for underwriting decisions—including ACORD forms, Statements of Value (SOVs), and loss runs—required meticulous, painstaking manual data extraction. This placed a heavy administrative burden on associate underwriters, which diverted their focus away from a key responsibility, critical risk evaluation.

Furthermore, they needed better visibility and accuracy of the data flowing between their core systems, including their Guidewire platform. The high volume of documents, potentially surpassing 400,000 per year even outside of CAT season peaks, exacerbated inefficiencies, which impacted turnaround times and the teams’ ability to swiftly respond to policyholder needs during their most vulnerable moments. In this present state, the carrier’s existing system struggled to keep up with typical daily volumes, let alone the overwhelming surge experienced during the June through November hurricane season.

The Solution

Faced with rising demands and an overburdened workforce, the insurer turned to Roots for a dedicated partner to transform their underwriting and claims operations. Roots proposed an innovative advanced AI Agent platform powered by Roots’ proprietary generative AI model specifically trained for the complex terminology and processes common in the insurance industry,  InsurGPT™. This was more than just the introduction of a new tool; it signaled a new level of operational efficiency.

The Roots’ solution was delivered in phases to address the insurer’s most pressing needs systematically. 

Phase one focused on Claims Indexing. Roots’ AI agents were seamlessly integrated to pull incoming emails via API. They then automatically extracted basic claim data from a multitude of document types and intelligently classified these documents using an AI model, and classification categories tailored to the insurer’s specific business needs. This rapidly applied much-needed consistency, standardizing document types across both adjusters and administrative teams. All extracted data and classification types were converted into a structured JSON format and were seamlessly transmitted back to key systems, significantly enhancing data flow and utility within their Guidewire environment.

Phase two tackled Commercial Underwriting Processing. As with claims indexing, Roots’ AI Agents rapidly automated the entire lifecycle of underwriting submission emails including multiple attachments that typically accompanied these forms. The Roots AI Agent intelligently handled submissions intake attachments, accurately identified complex forms, ACORD forms and schedules of values (SOVs), and extracted the needed data with a high-level of precision. The system even performed automated 3rd-party entity searches against the Division of Corporations database, further reducing manual underwriting work. As with claims, all underwriting data was converted to JSON format and fed back to key systems, transforming previously unstructured, cumbersome data into actionable, strategic insights.

The impact was significant, and the insurer saw a:

• 70%+ increase for claims document processing capacity
• 60%+ increase in underwriting document processing capacity
• Significant reduction in manual document processing activities
• Virtual elimination of submissions clearance backlogs
• Improvement in the consistency and accuracy of classification activities

Working with Roots empowered the insurer’s teams, automating numerous critical labor-intensive processes to shift their focus from tedious data entry to more complex decision-making and policyholder service activities. The insurer could now confidently pursue business growth without a corresponding increase in headcount, all while significantly enhancing operational speed, data accuracy, and the overall quality of their decision-making.