There’s a new buzz word that’s starting to make itself felt. That buzzword is “The Internet of Things” but what does it mean? And more importantly what impact is the “Internet of Things” going to have on insurance.
The Internet of Things
The Internet of Things is a wide term which really covers the idea that you might want to (and be able to) connect non-computing devices to the internet. For example; you might want to be able to put your dinner in the oven before you leave home but you might need to be able to control when the dinner starts cooking based on when your last meeting ends.
The internet of things would allow you to put a chip (computer not potato) in the oven. This would connect it to the internet. You could then use your mobile phone to let this chip know that your meeting was over and it was time to start cooking the coq au vin.
In essence the internet of things is anything at all which might be connected to the internet. It will lead to “smart devices” in a similar way that the last decade led to “smart phones”.
How Might This Impact Insurers?
Imagine that you could connect someone’s fridge to the internet and then monitor the contents. If you were a health insurer; you could then offer the owner benefits if they kept their fridge stocked with healthy fruit and vegetables and perhaps penalties if their fridge overflows with lard and burgers.
Carpets could be programmed to detect if an elderly relative falls on the floor or if a burglar has stepped into your hallway. Insurers could take this data and reduce senior-care insurance costs or home insurance premiums.
Thermostats, window controls, etc. could all be plugged into the internet of things. Careful control over the humidity or moisture content in the air of a house can prevent dry rot and other issues. Again premiums could be tweaked for householders that agreed to meet and abide by their insurers standards.
Once again, health insurers should be salivating at the prospect of exercise equipment that reports on the user’s commitment. Failing to do the right amount of work on the treadmill or exercise bike in a certain period could have significant impacts on a consumer’s premiums.
There are dozens of other potential uses for the internet of things. What the real implication for insurers is is simple. It’s the idea that there will be much more data available for insurers to access risk on. This should lead to ever more personalized products and better value (and profitability) from policies as a whole.
A recent report in the US; The State of Insurance Fraud Technology revealed some worrying numbers. In particular, only 14% of insurers are using advanced technological fraud detection techniques. Given the rise in sophisticated insurance fraud rings – this seems a little disturbing. However, it may just be that the industry isn’t aware of what can be done. With that in mind we’ve put together a quick guide to the kind of technology that can combat insurance fraud.
Whilst this might sound like a Search Engine Optimization technique; it’s not. Link Analysis is the use of modelling software to map complex relationships between large sets of insurance data. It can go beyond the borders of the insurer and link to data from solicitors, hospitals, garages, etc. to see if there are patterns of behaviour that indicate fraudulent behaviour.
This is software the takes previously identified patterns of behaviour and extrapolates them in real-time against insurance data. The idea is that predictive modelling lets you know when fraud is taking place, preventing payments to the fraudulent parties and allowing insurers to take a long, hard look at suspicious claims as and when they arise.
Not all data is neat. Much of the data that insurers collect, for example call data from call centres, is unstructured. Being able to analyse that data for patterns requires the use of text mining; this provides the ability to wade through huge volumes of unstructured information and still identify useful data or patterns within that data.
Massive data sets can quickly become overwhelming even to the most skilled analytical teams. Software analysis is generally very good at recognizing patterns of behaviour, which have already been identified at some point in the past. It’s not as good at making the intellectual leap to identify new patterns of suspicious behaviour. Data visualization allows huge complex data sets to be represented in visual format; this allows analysts to search for patterns that simply wouldn’t be visible in the original raw data sets.
Geographic Data Mapping
Geographic data sets are often hard to correlate without mapping them on a visual overlay. Insurance fraudsters often have to organize in tight areas of geography in order to get a consistent approach to making fraudulent claims. The ability to see trends by geography – gives insurers an effective means of finding these rings of fraudsters early and cutting them off before they cause too much economic damage.
Insurers have never had as many tools at their disposal to combat fraud as they do today. Investment in these tools needs to increase in order for them to reach their maximum effectiveness but the message should be clear; those that do invest are likely to find that the 2.1 billion GBP lost each year in fraud by the British Insurance Industry can be dramatically reduced.