Insurance Carrier: Claims Leakage & Processing Efficiency Improvement with Deep Standardization, Analytics & Automation

Client situation

Claims Leakage & Processing Cost Reduction

How Deep-Data Standardization, Super KPIs™ and Automation Generated Near-Term Benefits for the C-Suite

A regional leader in personal lines property and casualty (P&C) insurance consistently led its peer group in policyholder experience and service performance. Based on the market share in its largest product category—auto insurance—this insurer is positioned in the top 50 auto carriers within the U.S. market.

However, both its expense and loss ratios were routinely higher than competitors’ and stubbornly resisted management’s persistent improvement efforts. Technology implementations alone weren’t delivering the promised cost or productivity benefits.

The company had always believed that the latest core technologies offered the best hope for improving both efficiency and effectiveness.

Their claims-management system was state-of-the-art. And an app powered by artificial intelligence (AI) produced damage estimates directly from customers’ uploaded photos.

The company’s internal robotic process automation (RPA) team had implemented a half dozen bots. But these hadn’t come close to capturing the major efficiency, scalability, and profitability benefits that could be generated from successful, organization-wide, RPA implementation.

Operating costs weren’t budging.

Management was seeking a new strategic process-improvement partner—one that would deliver benefits and further prepare their organization for deep automation. They contacted The Lab after a peer shared videos from The Lab’s YouTube channel demonstrating insurance-industry automations previously implemented for similar processes.

Project Approach – A process-first, benefits-driven strategy

The claims improvement initiative began with a seven-week Phase I Analysis & Discovery effort that mapped end-to-end business processes and the supporting system data from FNOL through final payment and subrogation. Objectives included documenting the improvement opportunities, quantifying the benefits ranges, and developing a detailed strategic implementation plan that reflected the client’s needs.

Key Findings from Phase I Analysis

The Phase I analysis of the client’s organization revealed large cost-reduction opportunities that could be realized in the near term. Most of the savings—60 percent—were concentrated in reducing claims losses and increasing subrogation recoveries, i.e., claims leakage. An additional 40 percent were available from increased processing efficiency. One of the major takeaways from this analysis was the absence of simple, data-driven KPIs to monitor claims performance—all the way down to individual performers.

Major process improvements were uncovered throughout end-to-end claims processing,
from FNOL through subrogation. Low performing adjusters & intake data quality was in the spotlight.

Representative Examples:

1. Claims intake/FNOL errors – Many aspects of claims-intake data were under-standardized, outdated, and “un-automate-able” This resulted in subjective decisions by adjusters, policyholder confusion, and extensive downstream rework. Using data science and process mapping methods, The Lab pinpointed two issues causing over half of these FNOL errors:

• Incorrect cause of loss – Outdated, unclear definitions caused confusion for both FNOL reps and policyholders.
• Cumbersome adjuster assignment – A confusingly unusable array of over 40 available assignment “paths” and claims-management-system workflows was reducible to just eight.

2. Claims severity variance – Data-driven analysis of claims paid over the previous two years revealed large employee-level variance in payments for comparable claims. This variance resulted in numerous inaccurate “over-payments” each year, largely concentrated among lower-productivity employees. The Lab worked with top adjuster managers to define “red flags” warning of potential for higher-than-needed payments.

Examples:

• Claims open for 10 or more days, with no file note
• Full replacements (vs. partial replacements)
• Above-average inbound call volumes (by desk adjuster)

Leakage analysis revealed that not only did the low-performing claims adjusters pay fewer claims, but they also overpaid each claim—by an amount equal to just over one-half of the total expected loss.

3. Subrogation referral standards – Roughly half of all inbound subrogation requests were dropped because the forms were not in good order, or NIGO, and no error tracking or quality control (QC) review was in place. The missing data was available but required a search effort from the subrogation processor.

Examples:

• Missing adverse carrier information
• Incorrect claimant contact information
• Insufficient damage details
• Application of depreciation

Phase II Implementation: Claims Overpayment Reduction, Automation, and Strategic “Super KPIs”

At the conclusion of the Phase I work effort, The Lab organized the standardization, analytics, and automation improvements into a “modular” roadmap for implementation.

Major implementation objectives included standardization and automation of claims processing end-to-end for major labor savings and policyholder-experience gains. The improvement modules (workplans) included:

• Operational “Super KPI” analytics reporting
• FNOL intake & assignment automation
• Claims processing automation
• Claims accuracy management
• Subrogation referral and research automation

Solution 1: FNOL Data and Process Standardization

The Lab began simultaneous improvement efforts in claims intake (FNOL) and claims processing. The goal of the FNOL improvement effort was to immediately consolidate intake—the source of much downstream rework.

Standardization

The goal of claims processing improvement was twofold: 1) Reduce the average processing variance for similar claims (which had been measured at >7x), and 2) Investigate the payment standards and other sources of variation across the organization. Issues addressed included:

Under-standardized FNOL calls

A lack of standardization during FNOL calls created rework that could persist throughout the life of the claim:

• Incomplete, incorrect claim details
• 80% of claims required follow-ups with claimants: FNOL and adjusters
• Data spread across multiple systems
• Avoidable research for adjusters

De-centralized, limited-visibility data

The same FNOL call notes were entered into multiple systems, resulting in:

• Duplicate, manual data entry: FNOL
• Avoidable data-entry errors: FNOL
• Inability to see claimant notes: adjusters
• Redundant questions for claimants: adjusters

Almost immediately, The Lab began to break down these divisions between FNOL and adjusters. Temporary provisions were made for data sharing. System access was granted to both FNOL and adjusters. Forms were modified to standardize intake capture. Select fields were equipped with if/then logic and “lockout” options to ensure complete data collection.

The implementation objective was to narrow the information collected and keep it in a single, unified format.

Solution 2: Claims Intelligent Automation (RPA, AI)

As the above mentioned data was consolidated, preparation for automation began. The format of the new, digital FNOL collection form was similar to the old, “manual” form, and it was modelled on ideal workarounds that staff used for the existing core technology’s limitations. Loss causes were standardized and arranged in a hierarchy. An RPA bot for data QC was developed to correct common errors and validate details. Another bot entered the cleaned FNOL data into various systems: Claims processing (Guidewire) and the CRM (Salesforce).

Claims adjusters

The Lab worked alongside claims managers to identify new standards for data validation, consolidation, and standardization of claims intake errors. Once simplified, these options were committed to automation-enabling dropdown menus and a series of downstream bots to process the work.

Subrogation

Simple adjustments to data intake forms enabled automation, and lockouts effectively error proofed these upstream in the process. RPA bots monitored intake quality and generated real time alerts for errors.

Additional automation efforts were undertaken for:

FNOL data intake, QC, and entry

The heavily manual entry of claim-intake data into the claims processing system (Guidewire) was automated by upgrading digital forms that bots could pick, QC, triage, and resolve to ensure that critical, required information was collected at FNOL for all claim types.

Claim data and documentation indexing

Claim information transfer was automated with RPA from the intake process to claims processing
systems and other archival imaging systems. Any missing information was auto-researched and compiled by a bot for the adjuster to review. Distribution of the approved, auto-generated information list to the claimant was also automated.

Claim reserves and payment calculation

Initial reserve ranges and loss-payment amounts were auto-generated at each appropriate stage, based on claims-readiness data and historical performance.

• Reserves made out-of-range were flagged by a bot for escalation, validation.
• Loss payments made out-of-final-reserve-range were tracked by a bot & escalated for detailed review.

Claim-status notification and actions

Newly installed automation proactively updated the claimant regarding the status of their claims, what to expect, and whom to contact with questions. This automation logged communication into appropriate systems and identified next-step actions for the adjuster as well.

Solution 3: Upgraded Claims “Super KPIs™”

To sustain these improvements, supporting KPIs—aligned to strategic objectives—were installed.

Management was eager for new business measurements. Within eight weeks, The Lab rapidly installed its standardized insurance “Super KPI” reporting suite. Twenty vital KPIs were compiled daily in a “top-down” executive-summary view. Product, location, policyholder, adjuster, and other applicable data drill-downs were made available for each KPI.

The primary focus of these modernized analytics reports was to provide better data—in real
time—to manage claims-payment accuracy. This required a three-step approach:

1. Apples-to-apples claims identification – The Lab worked alongside underwriting and actuarial staff to standardize parameters for matching similar claims to one another for comparison. This included several facets:

• Cause of loss
• Location (down to the census tract, where possible)
• Date of loss
• Vehicle make/model (auto-only)
• Exposure type (e.g., PIP, bodily injury, etc.)
• Claims attributes (e.g., CAT/non-CAT, total loss, etc.)

2. Historical claims data mining (automated) – After standard parameters to match similar claims were identified, automated analysis was run to set a historical range of reserves and payments for each comparable claim type. These ranges were set each week and stored in a dedicated database table.

3. Real-time loss reserve recommendations & monitoring – Automation was put in place to match each new claim (after FNOL) to similar historical claims and send the assigned adjuster a detailed report to assist with setting accurate loss reserves. Additional reports were created to monitor adherence to reserve recommendations at the individual-adjuster level.

Implementation Benefits

The implementation work was completed within six months and delivered major financial, operational, and policyholder-experience gains. The Lab’s templates and knowledge base of standardization improvements, automation use cases, and KPI dashboards kept the transformation effort moving swiftly.

Standardization and automation-readiness improvements in FNOL data intake delivered immediate results.

With more complete information up front, “downstream” teams reported a 70-percent reduction in NIGO claims and resulting rework. AI-enabled automation generated claims files with preliminary adjudication notes and next-step actions. Subrogation and salvage opportunities were flagged by bots and forwarded to the relevant teams.

Easier processing also benefited policyholders. Avoidable FNOL follow-up calls declined by more than two-thirds. Automations proactively updated claimants on the status of their claims, what to expect next—and when. These notifications pre-empted most calls from claimants asking for status updates. End-to-end claim-processing cycle time—from FNOL to claim closing—was slashed nearly in half.

But what excited executive leadership the most was the visibility delivered by The Lab’s Super KPIs. These provided unprecedented insight into enterprise-wide performance. Productivity. Quality. Service levels. All reported in an executive-friendly summary—understandable at a glance—but with the ability to drill down to the individual employee level. These enabled continuous, data-driven improvement to sustain and expand the project benefits.

Book your demo today

Rapidly accelerate—and monetize—your carrier’s strategic initiatives, including process improvement, org design, intelligent automation, analytics, AI, and more.

To book your 30-minute screensharing demo, simply call (201) 526-1200 or email info@thelabconsulting.com