Artificial intelligence – Assisting insurance
Posted on: 30 August 2018
By Adam Rates, head of IT strategy and architecture
Artificial Intelligence (AI), once the stuff of science fiction, powers a whole host of things which are now part of everyday life. Just think of Siri, Cortana or Alexa to start with.
AI is a term that covers a wide range of technologies and capabilities which can be of huge benefit to businesses. This is equally true for the insurance industry, with many insurers exploring both AI and an associated technology - robotics - for functions, from claims handling to fraud detection.
The terms AI and robotics are not interchangeable. Robotics uses information technology for automation and for executing rules-based processes without any need to understand the steps being followed. AI is inherently more sophisticated and commonly involves training a computer on an activity’s outcome. The difference is that the computer programme “learns” and improves at the outcome the more it does, hence the term ‘intelligence’.
The use of chatbots is a key example of AI in the insurance industry. These computer programmes mimic human interaction and are able to answer common questions like “am I covered for…?” or “what is the excess on…?” The advantage is that they are easily available available 24/7 and can deal with multiple customers simultaneously.
As the chatbots become more sophisticated, we will, in time, see them dealing with more complex tasks, like settling simple claims. Insurers are already reporting successful pilots of this technology, with very positive customer feedback and impressive efficiency gains.
Fraud detection is another area of interest. Complex algorithms can analyse patterns, such as an individual’s spending data and identify anomalies suggestive of fraudulent activity. AI can also work analysing images used in the claims process, such as photographs of damaged vehicles. After being fed thousands of images, over time AI can develop the capability of recognising where an image has been doctored. Having confirmed authenticity, AI subsequently assesses the damage and could estimate the repair based on the damage seen.
AI promises a good deal. It will take time to develop and should be viewed as enhancing rather than replacing the human. AI can take away the mundane to leave a person to do the complex work. For example, brokers can use AI to assist in analysing market options but ultimately their role relies on an understanding of, and relationship with the customer. Machines are a long way from this!
This article was originally published in Modern Insurance Magazine (Issue 33). It may not be replicated in any other publication.