The insurance industry has had data at the core of its business for long – long enough to have taken the pole position in adopting data science and analytics. However, it is the banking industry that spearheaded the use of data science for commercial applications. Subsequently, industries such as retail, healthcare, CPG, etc. joined the analytics wave finding a wide array of applications of analytics. But, the insurance industry lagged behind in this context.
The insurance companies continued to stay with the traditional approaches to underwriting and actuarial processes and policies, so as to not rock the boat. For example, underwriting was traditionally based on a set of business rules, which could range from simple, to complex, to intricate. These rules evolved over decades, with business knowledge being passed from one generation of underwriters to the next. These rules were customized to evolving business scenarios and regulatory frameworks. They would be prioritized and picked by underwriters per situation and subjectivity of the carrier organization.
While this was a common scenario among insurance companies, banks were making great strides in moving away from rule-based customer risk assessment to purely data-driven approaches, using various statistical models – decision trees or logistic regression-based approaches on clustered data. In the last two decades, this became a norm in the banking and lending industry, which led to better risk-based pricing.
Having missed out on the first analytics wave, insurance companies had to play a significant catch-up and they did that with vigour. Today, they’ve implemented analytics solutions in areas such as customer acquisition, underwriting, actuary, policy persistency, claims, and fraud detection.
In fact, insurance companies have now moved on to the next stage, termed ‘Analytics 3.0’, where some “smart” carriers are going beyond just performing statistical modelling on historical/customer data.
In this context, we’re seeing two key themes emerge:
1. Augmenting internal data by tapping external data sources, both free and paid. These external data help them close key gaps in their data and provide valuable additional signals.
2. Using a wider variety of internal unstructured data – text, speech, image, audio, video – to drive enhance existing processes.
Related to the first theme, we’ve worked with P&C carriers who are making very good use of external data, primarily for the Small & Medium Commercial line. The data enrichment from external sources (in excess of 50 plus sources) augments application pre-filling, straight through processing, underwriting and claims processing. This helps cut down the manual effort and timelines to prefill applications, verify applicant-provided details, process applications and better underwrite risk. Similarly, Life Insurance carriers are using medical records and prescription data, for underwriting and straight-through processing purposes.
On the second theme, we’ve enabled insurance companies to leverage AI in a wide variety of innovative applications. For example, we have helped analyze drone images of rooftops to quantify data, used property inspection notes to track and monitor risks, and determined locational risks using geo-tagged/satellite imagery.
From these, it is quite evident that the insurance industry has now gathered significant steam in the innovative adoption of AI and Analytics.Tags: AI Insurance Analytics Insurance Analytics Trends