Insurance fraudsters are increasingly being caught out by technology, as the story of a
Florida woman caught faking an injury on CCTV showed recently.
But now, sophisticated techniques to analyse policyholders’ social media networks are being employed by insurance companies to combat fakers.
Though statistical analysis itself is not new, technology and the rise of social media have created the perfect storm in which companies can harness machine learning for more in-depth analysis of data, Michael Elliott, senior director of knowledge resources for The Institutes, told Insurance Business.
Social Network Analysis (SNA) uses data gathered from people’s social media profiles to look at their networks, behaviours, and even the sentiments they express through language.
“It’s more than what you first think of – Facebook and Twitter – you can apply Social Network Analysis to linked web pages, to telephone calls, it’s really a network,” Elliott explained.
Text Mining, which can be used alongside SNA, involves taking text and turning it into numbers, for example, through looking at what words are mentioned and how often.
The techniques are still relatively new, but are being used for various purposes within insurance, Elliott explained.
“For example, an underwriter would like to know very soon after a claim is reported, whether that claim is likely to become complex and a high severity claim,” he said of the use among underwriters.
Crucially, Social Network Analysis and Text Mining can also help project whether a claim is likely to be fraudulent.
The potential for these techniques in the insurance world is huge, according to Elliott, who predicts that there will be particular growth in the next few years in the claims and underwriting space.
While the end goal is for the systems to be able to automate processes, there will still be a role for underwriters who will need to oversee the process, and, ultimately, analytics can help insurers to keep up with the competition.
“You’re able to segment your market better and price your risks better, so insurers need to do it in order to stay up with their competitors, but there’s still a role for underwriters. It helps the underwriters do their jobs, but the underwriters still need to be looking at what the model’s doing.”
Related stories:
Far Out Friday: Head scratching workers’ comp claim foiled
The problem with insuring car-makers
.jpg)
