Global insurance company MetLife is taking the Internet of Things (IoT) and machine learning very seriously, despite still being in the early stages of implementing and exploiting the technology, according to MetLife VP of enterprise analytics Malene Haxholdt.
Haxholdt told V3 at a SAS Global Forum media event in North Carolina that the IoT has the potential to help the company develop new products, but that work in this area is still very much at an “exploratory” stage.
“IoT is definitely being considered as something you have to take seriously in this business. It’s not an ‘Oh it’s going to go away’ type of conversation. That data is not going away, so you have to take it seriously and figure out if there is a use for it,” she said.
Haxholdt explained that the IoT can play a part in SAS' digital transformation efforts, enabling it to communicate better with customers and use real-time data analytics to improve efficiency in its underwriting process.
“It could allow us to do it much faster yet still be very precise because of the data we’d have to support it,” she said.
But Haxholdt believes that the accuracy of data from devices such as FitBits, which could have an impact on health insurance, is still not reliable enough for the company to exploit.
“We’re still at a stage where there are questions about data accuracy and privacy. We could get a lot of data but do you have enough? Do you have it in a structured way to do some trustworthy analytics on it?” she said.
Machine learning could play a big role in enabling MetLife to harness IoT data. "We’re exploring a lot of our access to unstructured data and we’re using machine learning for that,” explained Haxholdt.
This machine learning is being developed in-house by MetLife data scientists, but the company has also partnered with external organisations in this area.
One potential application for machine learning is with speech analytics in the company's call centre.
Machine learning, like the IoT, crops up more and more in conversations, according to Haxholdt. “It is definitely there, it’s just about how you define it. We define it as deep learning using algorithms that go deeper and iteratively through data,” she said.
“If you look at data from cars and Fitbits and so on, it is going to depend on being able to do machine learning. And when you think about digital interaction, where you need intelligent decision-making in a real-time environment, you’re going to have to rely on machine learning to do it faster with a sense of accuracy.”
Want to hear about some real use cases for the IoT? Come to V3 sister site Computing's Internet of Things Business Summit on 12 May in London. It's free for end users.
Could be used for everything from search-and-rescue robots to wearable tech
Don't require the rare material being mined from the mountains of South America
IBM hopes that its new tool will avoid bias in artificial intelligence
Found by calculating the strength of the material deep inside the crust of neutron stars