How CDOs Should Look to Handle Their Pools of Data
"Along every link in the value chain—from underwriting, sales, and distribution through to claims management, recoveries, and even product design—insurance companies are looking at ways data transformation will empower more strategic initiatives," says Alex Johnson, Head of Insurance Solutions at Quantexa.
These efforts aim to improve combined ratios and re-engineer operations to become more efficient and personalised. The motivation is clear: successful implementation can lead to substantial improvements in financial performance, even within a tightly constrained economic and regulatory environment.
A 360-Degree Customer View – Is It Worth It?
"Investments in data and analytics are essential—everyone is doing it. However, insurance companies face a vast, fast-flowing river of data fed from all sides by internal and external sources. This temptation to siphon off data and build individual platforms and services is understandable. However, according to Gartner, 85% of AI projects fail mainly due to poor data quality. This shows that the current approach exacerbates common issues around data quality, governance, and usability, resulting in missed opportunities and increased risk," comments Alex.
Poor data quality costs businesses between US$10mn and US$15mn annually, which will increase as business and information systems become more complex.
Creating Locks in the Data River
Alex Johnson suggests a more sophisticated approach to managing data quality: "Create locks along the river, slowing down and making sense of the data as part of the overall system, not in siloes. True data value comes from what Gartner calls 'Decision Intelligence'—a holistic, unified analytical view centered around every applicant, customer, claimant, third party, supplier, and risk across the organisation. While this concept sounds great in theory, only 18% of insurers are currently optimising their data use for competitive advantage."
"CDOs, CTOs, and their peers are understandably cynical of being sold another 'silver bullet,' especially when failures are often attributed to 'data quality issues.' Good data management is foundational to every product, service, and project undertaken."
Good Data Management is Foundational
"For insurance companies, resolving both structured and unstructured data quality issues at scale is crucial. This capability, called Entity Resolution (ER), involves resolving multiple labels for individuals, products, or other insurable objects into a single entity and analysing relationships among these entities," he says.
Typically, ER requires a data transformation exercise to format data before matching. Traditionally, this is where processes slow down or fail, taking years to ingest and match data accurately. However, the latest dynamic ER versions can handle data in various formats and start creating a single unified view.
Building Relationships and Insights
"Once data wrangling is complete and a unified, trusted data foundation is created, teams can start revealing relationships and insights across multiple business areas. This is impossible when data is siloed."
Context is crucial for informed decisions, reflecting structural and transactional relationships across different individuals, organisations, companies, and claims an insurance firm tracks. ER parses, cleans, and normalises data using sophisticated AI and machine learning models to reliably identify entities. This clustering and labelling of records are more effective than traditional record-to-record matching used by MDM systems.
"An insurer might receive an application from a business seeking liability insurance. The business may have been previously insured under another line of business, with a bad loss history due to employee misconduct. This information is vital for assessing new applications. Insurers excel at top-level segmentation and risk assessment but often lack expertise in evaluating customers based on their comprehensive risk and loss history."
Leveraging External Data
"New legislation, such as the Financial Data Access Framework (FIDA), allows insurance companies to access a broader range of external data. High-quality entity resolution can link internal data with high-value external data, such as corporate registry information, which was previously difficult to match reliably."
"CDOs are under immense pressure to deliver actionable results quickly. Their success depends on ensuring their organisation’s data is accurate, fit for purpose, and accessible to those who need it. All AI models require these conditions. Implementing models and generative AI solutions without ensuring data quality, governance, and usability first can feel like pushing water uphill. By addressing these fundamentals across the whole data river, CDOs can achieve Decision Intelligence across the entire organisation."
**************
Make sure you check out the latest industry news and insights at InsurTech Digital and also sign up to our global conference series - FinTech LIVE 2024
**************