As with so many other industries, insurance is being disrupted on a digital level, and data is both the cause and the cure for that disruption. The availability of new data sources has created new opportunities to reduce risk and exposure as well as creating tailored products based on customer profiles. Simultaneously, however, this has caused complexity and the sheer scale of data has overwhelmed many organisations.
Insurance firms need to become more adept at extracting value from unstructured and real-time data, as well as utilising it for predictive analytics and machine learning. Once they have a handle on this, the idea of risk within insurance will be transformed and the way brokers, underwriters and claim handlers operate will be changed forever.
Boosting productivity and faster application processing
Many insurers sell policies through agents. To prepare for sales calls (or to answer questions from prospects during those calls) those agents may need to look up details across multiple disjointed platforms or applications. This takes time and, due to high volumes of data to be considered for the smartest, most effective pitch, velocity decreases. A more advanced data platform should instead store data from a multitude of sources – including insurance IOT – in a ‘data lake’.
This permits a single lookup, without requiring multiple individual queries across different unrelated storage platforms. Agents can prepare themselves more thoroughly and much faster. Insurance companies can also use the same type of single view to understand which agents are most productive at selling their products, offering incentives that promote top performers or de-certifying the chronically unproductive; add machine learning to this and all can be done with minimal workforce investment.
Once customers have the need for new insurance coverage and apply for a new policy, the agent and / or underwriter needs to process the application documents. This can be a time consuming process and error-prone in the data analysis leading to potential premium leakage. Speed is important, but so too is accuracy. With a data architecture that facilitates multi-tenant processing, real-time data and machine learning, the same data set can be analysed without compromising speed or accuracy, insights can be utilised and recommendations quickly adopted.
Valuable data that solves discrepancies saves money and reduces fraud
Insurance firms often struggle to gain value from their data. Structured data systems and architecture have often locked data into a source. By utilising the unstructured data that surrounds a claim, however, a more complete picture of the claim can be built and machine learning can be applied to tap into this broader spectrum of data sources. For example, in a car accident, machine learning could enable the claims adjuster to tap into context such as the weather conditions, dash cam videos or data gathered by the city’s smart transportation system.
Evaluating claims data and evaluating for cases of subrogation can also be an issue in the insurance sector. By applying text analytics to the unstructured data, in the form of claims notes and diaries, it is much easier to identify areas of missed subrogation or recovery. For some organisations, this can amount to millions of pounds, a significant sum that could be better used for gleaning insights from that data, rather than just correcting existing problems.
With detailed information around a claim, coming from different sources and automatically ingested into the data analysis platform, it is far easier to identify possible fraud and reduce risk when validating legitimate claims. After the claim has been evaluated, machine learning could also aid in claims payment optimisation to reduce the chance of claims leakage by the carrier. In turn, all this data could be fed into the carrier’s predictive model to identify the types of claims that leak versus those that do not.
Ultimately, the ability to tap into a wealth of data and then apply machine learning in the insurance industry will benefit both the claimants and the insurance firms. Evaluating risk has always been somewhat of a guessing game, but with data insights at their fingertips and an intelligent system that can act on this data, insurance firms can make much smarter underwriting decisions both in the short and long term. With an optimised risk appetite and more informed pricing decisions, machine learning has consequently driven down claims costs and, as a result, premiums for the more risk averse. Customers are much more likely to have a better understanding of what their policies cover as precision pricing is more accurately aligned with the risk.
The above article was an opinion piece written by Cindy Maike, GM Insurance, Hortonworks. The views expressed within the article are not necessarily those of Insurance Business.