Where are we now when it comes to AI?
2017 was well and truly the year than AI hit the mass-markets and grabbed the attention of consumers - both young and old. All across the western world, Google Homes and Amazon Echoes of all shapes and sizes were welcomed into consumer homes with open arms. For personal use, these assistants can tell you the time, keep your appointments, and search the web when you have your hands full – and their capabilities growing by the day with each new input and request. Many are starting to ask what they won’t be able to do in the future.
In the business world, the story is similar, if not quite as visible yet. More and more companies, especially within the finance sector, are exploring what AI can do for their business, to either plaudits or scepticism from employees and the general public.
Some see the advancing technology as a useful progression, with the potential to free up their staff to tackle more complex problems and create new innovative solutions. With this view, AI systems are seen as “co-bots”, working in tandem with humans to achieve better, faster and more reliable results. Others, though, see a threat to their jobs and the economy with every new development.
One recent study found that around half of the activities people are paid to do globally were at risk for automation using currently demonstrated technologies, and in around 60% occupations at least one-third of activities could also face automation.
Taking this further, by 2030, it is estimated that up to 800 million individuals will face displacement by automation and need to find new jobs. Workers in a multitude of industries may need to learn new skills and switch occupational categories in the years to come, risking intense upheaval and increased pressure on both public and private solutions.
How is AI and automation impacting the insurance industry today?
According to a survey done by Tech Emergence using data from the National Association of Insurance Commissioners (2016), the most popular AI applications for insurance industry leaders are:
1. Chat-bots / AI assistants
Chat-bots aren’t new, but their skill sets are increasingly improving. They are now being used to handle customer complaints and process simple transactions, with on-the-job training to help them continue to “learn” and move on to more complex requests.
In fact, Haven Life, a US insurance start-up, is now offering an alternative way to buy life insurance online: through its Facebook chatbot. Not only this, but the chatbot is designed to help educate US customers on the costs of life insurance, in a bid to tackle misconceptions about the insurance industry.
And Haven Life aren’t alone. Allianz have Allie (an online assistant available 24/7), Co-Op Banking Group have Mia (a bot which answers customer’s insurance queries), and Hanna is a chatbot which has been around since 2003 for the Swedish Social Insurance agency.
Will those working in customer service and sales roles within insurance companies still have a place as these bots become more intelligent and sophisticated?
2. Driver performance monitoring and real-time assessments
Especially within the automotive industry, some of the top insurance players are looking to big data and analytics to shed some light on customer behaviour to help them innovate their product and pricing offerings.
Some big insurers are looking for their customers to opt into data collection opportunities in exchange for reduced premiums and rates (think black-boxes in cars and video cameras to record driving behaviour). Not only is this a great opportunity for the consumer to save money, but the insurer gathers valuable information which can help develop a larger picture of driving habits and inform the entire spectrum of the business, from product development to claims management.
One large American player, Liberty Mutual, is also reportedly experimenting with a new app to help drivers involved in a car accident in real-time. After an accident, drivers can use the app to quickly assess the damage to their car using their smartphone camera. With the app’s AI component theoretically trained on thousands of images from car crashes, it could provide damage-specific repair cost estimates much faster than traditional insurer adjusters who currently asses auto damage.
3. Insurance market analysis
Machine learning algorithms are not only being used on an individual basis to save consumers money, they are also beginning to be applied to interpreting larger driver data to monitor market trends and identify key business opportunities.
Progressive Insurance is a clear example of this, as their telematics app, Snapshot, has reportedly collected 14 billion miles of driving data. Previously, this much data would have taken a vast amount of time to sift through with human analysts, or even basic computation methods and that time taken was found to hold back business prospects. With the use of machine learning, the company can interpret the data more clearly and quickly, enabling more accurate predictions to be made about the market.
Within this application, the need for skilled data scientists is clear - you can’t make use of vast data sets, even with machine learning, if no-one can formulate actionable business insights. This, then, provides a clear case study for machines and humans working together and generating new employment opportunities rather than the common notion that AI only destroys jobs.
So, will there still be a place for humans in insurance?
It seems unlikely that AI will displace humans in the insurance industry, especially in roles that require subjective judgement and experience for claims management and underwriting.
Even those working on the frontline, dealing with customer enquiries and renewals, will find AI does not threaten their jobs as much as it will enhance their current capabilities and save them time. For one, the capabilities of AI are still underdeveloped and will require time and teaching to develop the ability to perform certain tasks, as well as a human overseer to step in when the machine cannot develop the required outcome. Many insurance activities are nuanced and multi-layered, requiring judgement in hazy situations, and it might be both too time-intensive and expensive to develop AI systems that can recreate that judgement.
Analyst and management roles are fairly well protected from the emergence of AI too, as machines provide insights but not the decision-making capability (or authority!) to drive a business. The insurance industry will always need the trailblazers and thought leaders to highlight a path to success.
“When it comes to how insurers should prepare for the [AI] shift”, AI technologist Francesco Corea emphasizes that insurance companies “should be ready to engage intelligently with new types of data and adapting their models and infrastructures to fully embrace the potential of AI.”