Market commentators believe insurance is on the cusp of major change driven by new technology such as telematics and investment in insurtech. But are insurers embracing this? Ant Gould reports-¦
Last month saw publication of the Future of General Insurance report by Marketforce in conjunction with the CII, following a major market survey covering everything from innovation and disruption through Brexit to fraud. The survey tracked key trends, disruptions and forces keeping insurers awake at night; this article focuses on the findings concerning two issues in particular; the Internet of Things (IoT) and artificial intelligence (AI).
The IoT has the potential to make our lives, at home and at work, easier and safer by giving a voice to the appliances and infrastructure that make up our physical spaces. However, to date, there is no sign of the industry rushing out smart buildings and contents policies: in the next 12 months, just 15% and 16% respectively of those responding to the survey say they will have a dynamically-priced proposition for buildings and contents insurance, although this rises to 32% and 35% within two years. Interestingly, pet insurers are keen to use connected devices for dynamic pricing, with 71% planning to offer telematics-based pet insurance within five years. According to the survey, the numbers of insurers currently offering dynamic pricing based on pet wearables is small -- just 4% -- but this is set to rise to 7% within one year, 17% within two years and 71% within five.
Crucially, these types of IoT-enabled products are not just about reducing premiums and improving underwriting accuracy: the data generated can be analysed, anonymised where necessary, and shared with policyholders. Up to 85% agree that dynamic pricing could greatly improve customer loyalty by creating more frequent and meaningful touchpoints throughout the policy life cycle; 96% expect it to reduce risky behaviour.
RISE OF THE ROBOTS Robotic process automation (RPA) is computer software that organisations configure to capture and interpret the actions of existing business process applications, such as claims processing or customer support. Once the 'robot software' understands these, it can take over the running of them -- it is, in effect, software automating the use of other software -- and it uses that software far more efficiently, accurately and cost effectively than any human.
The game changer however, is not low cost automation of routine tasks but artificial intelligence (AI), although this a broad term that covers many subsets of data science, including:
- Machine learning -- the science and engineering of making machines that do not just follow human commands but can self-learn;
- Deep learning -- a type of machine learning that uses multilayered neural networks to learn;
- Cognitive computing -- the simulation of human thought process based on how the brain works, using natural language processing and machine learning to enable people and machines to interact more naturally.
AI will analyse the claims process to identify bottlenecks and streamline workflows to reduce claim resolution times, eliminating stress for the policyholder and costs for the insurer. Deep learning models, for example, can analyse images and IoT data flows to automatically categorise the severity of damage to vehicles or buildings, estimate repair costs and trigger remediation for swift resolution of the claim.
AI: READY TO SERVE
Insurers are also tapping into AI to improve the customer experience. LV has deployed a virtual assistant that leverages AI technology, including natural language understanding capability, to deliver a more personalised and effortless customer experience for its brokers. The virtual assistant can understand intent by engaging with brokers to quickly access the information they need, allowing them to self-serve around the clock.
Data-intensive underwriting departments could also be transformed by AI, as intelligent machines sift through vast datasets and highlight new and emerging risks. Studies suggest that underwriters spend 70% of their time performing low value tasks, such as searching, aggregating and selecting data, and only 30% on risk selection -- a ratio that could potentially be reversed by the application of AI and machine learning.
Little wonder then, that almost seven out of 10 of our respondents expect to use AI extensively for underwriting within five years. Yet with some insurers -- such as Swiss Re -- already deploying cognitive computing, those five years could see a widening gulf in underwriting performance, states the report, between those leveraging AI's continuous and high speed analysis of vast data flows and those left behind, drowning in an evergrowing ocean of data they cannot hope to analyse.
â The full report can be downloaded for free at: http://events.marketforce.eu.com/gireport16-cii