The use of advanced analytics - including how to deploy artificial intelligence (AI) and machine learning to good effect - is one of the most widely debated topics in the industry. Willis Towers Watson analysed how P&C insurers are assessing the impact of this phenomena on their future business.
The web-based survey took place between May and July this year and captured the views of 42 P&C insurers in Europe: 30 multi-line insurers; 9 personal lines of insurance; 2 commercial lines and 1 specialty line.
According to Chapman the overall impact of advanced analytics on the industry is two-fold.
"From a technical perspective you might think the focus of insurers is purely on building better models with advanced analytics. But actually they are seeing it as a way of enhancing customer service and improving various parts of it.
"So it's expanding analytics into different parts of the business as well as models."
Fundamentally, a greater understanding of risk drivers is always a target. According to Willis Towers Watson, a large proportion of surveyed insurers are targeting this as a key application for AI and machine learning in the future. The survey states 28% of surveyed insurers currently see AI and machine learning as a top application for better understanding risk drivers. However, asked about the potential in two years' time, and the number skyrockets to 90%.
Chapman says analytics has long been deployed to increasing understanding of risk drivers, but the increased availability of improved data has made it an even more interesting area for insurers.
"Risk drivers historically has been an area where a lot of analytics has focussed.
"What the result is really articulating is that the use of AI and machine learning can supplement existing techniques and enhance the basis to better understand risks."
Willis Towers Watson is not the first to arrive at this conclusion. Earlier this year, a whitepaper from French reinsurer Scor stated AI could overhaul the entire risk universe.
Chapman adds the use of such developments could become standard practice.
"Insurers will go through a process of investigating different machine learning techniques and the process of embedding them.
"It's going to be seen as a requirement to have some of these techniques in place."
Nearly three-quarters (74%) of surveyed insurers said the main goal for analytics is in providing a faster service. Equally, 62% stated they intend to introduce advanced analytics to provide faster and easier access to information internally.
Scor's whitepaper also cited the impact of analytics on the distribution of products, notably how it could lead to greater efficiencies in underwriting, claims processing, risk analysis and product development. As a result, re/insurers, theoretically, will have a greater understanding of customers' risks and underwrite more effectively.
Chapman states the results reflect insurers' desire to improve operational aspects.
"We saw insurers showing clear interest in AI and machine learning in the claims process. This means you can have a more effective triage process for claims, and can effectively process claims faster.
"Customers will get a quicker response, whereas you might currently have a claims underwriter looking at it first."
Other key takeaways from the survey revealed how insurers' top data sources for customer centricity will evolve in the next two years.
Internal customer data, customer surveys and customer interactions are currently seen as top data sources by 85%, 83% and 82% of surveyed insurers respectively.
However, the survey revealed the important role social media is set to play as a source of data for insurers. According to the Willis Towers Watson survey, 55% of surveyed insurers see it as a top data source in two years' time, compared to just 13% currently.
"There is a recognition that social media can offer further information and data sources; it's quite a rich data platform and what the 52% recognise is that there is potential to better their offerings," says Chapman.
"However there is a question of how do you practically use the data in social media - and is it acceptable for the customer?
"For example, it doesn't necessarily mean we're going to use social media profiles to make risk models. But it could give further insight in suggesting how insurers and customers could engage each other more through social media."
This article was originally published by InsuranceERM, 11 October 2018.
All Rights Reserved, Reprinted With Permission
