For many insurers, a roadmap to implementing analytics begins with a packaged solution that can deliver immediate value followed by the development of more customized capabilities as requirements are clarified and technology staff become familiar with the data, algorithms, and tools available.

Mitigating Inappropriate Payments
Predictive analytics can be readily applied to fraud mitigation. In particular, these techniques support special investigative units (SIUs) in the detection of data-based, suspicious features of a claim that may not be readily apparent to the human eye. Fraudsters are quick to adapt their methods as insurers find and cover areas of vulnerability. Large-scale analytics can help identify new indicators of hard and soft fraud so that SIUs can become more proactive and respond quickly and more appropriately to suspicious claims. With these methods, insurers are typically able to reduce claims costs by 3 to 5 percent, while some carriers are able to attain savings as great as 5 to 10 percent.

Optimizing Workflow
Strong organizations are able to adapt quickly to change. Regulatory shifts, caseload severity, and customer service demands are three types of change that impact the claim organization's effectiveness. Using optimization technology, insurers can model and test the possible impact of workflow and organizational changes before initiating them.

For example, it is possible to estimate the impact of switching from a generalist to specialist adjuster staffing model. Predictive analytics can help answer questions such as, "Will the increased skill level offset the additional staff and training costs? By how much?"

Total Transparency
Many insurers are "flying blind" while attempting to find the best avenues to improve their effectiveness. Loss costs and LAE measures retroactively track results, but there are few methods to flesh out what specifically is driving those outlays. Analytics, combined with good data preparation, proactively highlight the correlations between caseload and operational factors and performance in a systematic way.

More Applications for Predictive Analytics
There are many ways in which predictive analytics can be used not only to assist in perfecting the claims process but also to serve as a training tool for seasoned claims veterans.

Being alert: clarity. Analytical capabilities range from simple exploratory statistics and understanding key performance indicators to advanced data mining and predictive analysis. Organizations new to data analytics should first grasp and become proficient with the concept by creating a clear visualization of their claims handling operation and building alert systems to notify them when performance is above or below a normal range. The best place to start is benchmarking performance against industry norms.

Being astute: prediction. Claims organizations are largely focused on managing costs after services are delivered. Predictive analytics help organizations move from that reactive to a more proactive mode of management. For example, analytics can be applied at first notice of loss (FNOL) to estimate the severity of the claim and determine if it would benefit from more intensive case management or if it should be fast-tracked. Analytics can also be applied in physical damage claims to help guide claimants to the most cost-effective repair center. These tools can be automated and built into the insurance company's workflow or can be used to create diagnostic reports that enrich the adjuster's overall knowledge and perception of each claim.

Being bold: prescription. Savvy claims managers recognize the limits of traditionally retroactive strategies to handling claims and explore analytics and automation to proactively manage trouble earlier in the claim lifecycle. One of the most important challenges is to recognize the severity and complexity of a new claim quickly and then determine the best strategy to mitigate potential risks. In high-volume claim departments, there are often not enough resources or time to thoroughly evaluate all the relevant information. The most advanced types of analytics operations combine transparency and prediction with prescriptive goals to deliver recommendations specifically tailored to a given claim situation. These tools take business rules engines to the next level by referencing a much broader base of claim information, evaluating all the potential tradeoffs, and ranking possible courses of action by their predicted outcomes.

The use of predictive analytics may be opening a new chapter in claims adjusting skills. Analytics help claims leaders solve the challenge of making essential decisions about strategy and claims management in a complex environment with many moving parts. While most organizations have an effective operating model for claims, the potential for higher-order improvements resulting from solid analysis of data can give insurers a significant competitive and service edge.