Some of these innovations include:
- The introduction of the Personal Claims Service wherein 75% of claims are handled by a single known individual called the Claim handler.
- Increased specialisation of agents so that the claims handlers can manage the complexity of the process they specialise in.
It is important that as soon as a claim is raised, the insurer is able to predict its complexity and the likely settlement route. Thus, the client wanted to identify ways through which the above initiatives could be complemented using a claims triage and profiling system so that the most suitable Claims handler could be assigned as early as possible. A single named point of contact assures the customer of a clear and consistent journey together with an agreed set of expectations.
The triage decision needs to be accurate and happen in a sensitive customer-focused manner. While the customer is raising the claim, it needs to quickly determine the likely claim profile. Wipro’s approach based its solution on the knowledge that by integrating a 3600 view of claims with the 3600 view of the customer and their coverage, it will be possible to extract the patterns that explain the differences between claims’ characteristics. These patterns can be used to define ‘claim profiles’ that form the basis of the triage step so that each new claim is sent to a specific expert who has the specific competency to deal with the selected profile.
The difficulty centred on the fact that the relevant data lay in many places (thousands of word documents containing claims reports of initial investigators), was voluminous (thousands of excel workbooks detailing damages, associated repair estimates and subsequent actual costs), and was often in narrative form. Essentially, this presented what we call a ‘Big Data Analytics’ problem comprising terabytes of unstructured and structured data on customers, covers and claims never assimilated and analysed before.
Wipro developed a strategy based on Text Analytics techniques to extract signals from text and blended them with data from structured and semi-structured sources into a larger modelling data mart. Many more sophisticated patterns emerged out of this strategy and the insight thus obtained played a powerful role in achieving the required accuracy of claims’ profile predictions.
Wipro demonstrated how its larger ecosystem can facilitate such a complex project to deliver a comprehensive analytical view through:
- Advisory Services: Providing advice on the best technology to invest in so that this analytical process could be repeated, enhanced and updated
- Diverse Technology skills: Applying highly specialised, qualified staff with the high-value skills capable of executing the modelling techniques ranging from text mining to clustering to predictive modelling
- End-to-End ownership: Provisioning complementary skill sets including data integration, data management, modelling together to manage the large volume and variety of data
The insurer now benefits from an efficient claims settlement process powered by predictive analytics. The claims handling process now has the capability to control the variance in costs, durations and customer interactions arising from the complex range of settlement routes. This has resulted in a 33% faster claims resolution process. The claims handler (whose assignment is based on the analytics that Wipro delivered) is now fully informed of the predicted key metrics of the claims and is able to keep the customer duly informed. This has resulted in higher levels of customer satisfaction – 28% reduction in customer complaints, and a steady fall in attrition. Overall, the initiative has led to a 20% reduction in cost.