Articles - The evolution of predictive analytics in insurance

More than most industries, general insurers use data and analytics to run their business efficiently, sustain profitability, and create competitive advantages. Significant investments have been made over decades to collect, organise, and analyse the massive volume of data that insurers hold. Given the spend, hype, and promise of big data, machine learning, and AI, many are asking the question, “What impact is the investment having on the challenges of capturing business value?”

 By Michael T. Anderson, Director Data Strategy, Guidewire Software

 While the insurance industry is one that has always used data extensively, this has often only been regarding to pricing and risk. Nowadays, thanks to the use of machine learning and artificial intelligence, insurers are adopting predictive analytics in their claims processes, particularly when it comes to identifying fraudulent claims.

 However, integrating predictive analytics with business processes is not an entirely straightforward process. Prior to any modelling activities, data science teams, business teams and IT teams have to understand fully the business needs and related technology needed to deploy the models into their core systems. Otherwise, problems will arise when insurers move to operationalize their analytics and the expected business value is often never realised. Moreover, even artificial intelligence deep-learning techniques – that is, artificially intelligent systems capable of learning unsupervised by looking at unstructured data – may struggle to understand patterns they have never seen before. Depending on the insurer’s needs, and the nature of the data they are using (structured or unstructured) variable amounts of data are required to create models. Fundamental to the whole process is that is the data should be high quality. Only then will insurers be able to build an AI capable of handling the complexities of the largest claims, whilst expediating the process for those smaller and more easily automatable ones.

 Within the insurance industry there are clear variations in the ability of insurers to capture measurable value. Insurers whose analytical strategy is focused around the final objective - delivering business value - tend to execute more effectively and deliver more value when compared to organizations who are not aligned with their business counterparts or perhaps overly focused on highly ambitious, if not somewhat speculative, objectives.

 Successful strategists consider not only the data needed for modelling and model development but, early in the planning phase, they place an equal emphasis on knowing exactly how end users will interact with the model output in their core systems so that it becomes part of their natural workflow. Quite simply, when organisational resources are laser-focused on a known, solvable, quantifiable business problem, the execution strategy tends to be better defined.

 Such an approach could - and perhaps should - alter a project manager’s decisions as to which analytics platform will best serve their organisational needs. Insurance companies often make the mistake of adopting a modelling or data science centric approach and this results in challenges when it is time for them to operationalise models into their core systems. Get it right, however, and predictive analytics has a proven ability to help redesign age-old manual processes, augment decision making, improve pricing segmentation and risk analysis, improve workflow through exception management, and redefine customer interactions.

 Those insurers who have made a success of predictive analytics have reaped the benefits. Take the example of one US insurer who attributes their use of predictive analytics in turning around their private passenger motor book. In 2015, the company reported an underwriting loss of over $35 million in its motor business, which clearly was unsustainable. In order to turn this around, they needed a way to be more targeted in their risk selection and to reduce new business growth. By employing predictive analytics, the insurer developed frequency models and a private passenger underwriting index, launched in 2015. The index scored new business submissions to identify potentially unprofitable risks. The results were significant. By 2017, the insurer’s underwriting function reported a modest profit of $500,000, which rose to $27 million by the end of 2018; all of which the insurer attributed to the underwriting index and the more sophisticated rate plan they had been able to implement thanks to their use of predictive analytics.

 Through investment in data, analytics and business solutions, insurers are better placed to improve outcomes in a consistent and scientific way, without human bias. As a result, they are better placed than ever to realise significant improvements in loss ratio and reduced expenses, and to grow their business profitably.

Back to Index

Similar News to this Story

Cutting your retirement cloth to fit your pension
Our local tailor has decided to retire. Not something which is usually newsworthy. However, Tarcisio is 88 years old. And he is still not ready to g
Climate models and investing what is the issue
Join Senior Sustainable Investment Consultants Clare Keeffe and Jordan Griffiths as they discuss climate modelling, and the many myths that surround i
How everybody wins in a digitized insurance market
The transformative shift to digitization enhances efficiency and collaboration for commercial lines insurance, providing advantages for insurers, brok

Site Search

Exact   Any  

Latest Actuarial Jobs

Actuarial Login

 Jobseeker    Client
Reminder Logon

APA Sponsors

Actuarial Jobs & News Feeds

Jobs RSS News RSS


Be the first to contribute to our definitive actuarial reference forum. Built by actuaries for actuaries.