Thank you for Subscribing to CIO Applications Europe Weekly Brief
All rise AI
By Natalia Kozyura, Head of Innovation Center, FWD Insurance
Five years later, when I had joined Amazon, the industry was just beginning to tap on machine learning capabilities to make the appropriate movies and TV series recommendations for our customers. Today, literally every product team at big tech companies in the world is evaluating the use of AI during new feature development.
The financial industry is no exception. For insurers, AI is being utilized for specific use cases like warm lead generation, claims processing, and customer servicing. It can also be used to automate internal procedures such as simple underwriting, real-time claim processing, and real-time payouts.
I would expect that within the next two years, every team in an insurance company will be looking at ways to leverage on AI, much like what their counterparts in the big tech firms are doing.
AI a leading disruptor today
Indeed, Big Data and AI are emerging as key disruptors across all industries, particularly for financial institutions. This is unsurprising, given the large volume of data banks and insurers alike have access to, where machine learning technology can be applied to create models and processes to serve customers better.
One of the most significant areas where AI can be used is customer experience. In the insurance industry where products can be largely similar across companies, delivering a cutting-edge customer experience can be the decisive factor that swings a customer from Company A to Company B.
This is especially pertinent for the customers of tomorrow–the millennials. As they continue to accumulate wealth and live through technological breakthroughs, they will expect to enjoy better–again, this refers to automated processes and personalized service–experiences when they interact with an insurer. For example, instead of presenting cookie cutter information on products and services, insurers have to instead deliver an e-commerce experience that is easy, direct and customized to who they are.
Needless to say, this group of discerning customers will expect to be able to find the best products suitable for their needs on digital channels (on their mobile devices, of course) without any needless waiting or energy expanded on processes that have, for better or worse, existed for centuries.
One of the most significant areas where AI can be used is customer experience. In the insurance industry where products can be largely similar across companies, delivering a cutting-edge customer experience can be the decisive factor that swings a customer from Company A to Company B
Against this backdrop, companies will do well to devise on an effective AI strategy if they do not wish to miss out on the massive potential it offers.
For the uninitiated, I have a simple formula to recommend. One, think big picture and understand the real job-to-be-done; two, develop use cases that will support your strategy to offer premium customer experience; three, learn from others and leverage on technology. If need be, learn from tech giants and ask for their recommendations; finally, break the entire process into multiple phases. To start, aim to deliver a real business value to customers and business units, whether it is to the actuarial department or marketing team.
Access to data is key to building AI
To start with the technical implementation of AI insurers and other financial firms need to leverage big data. The key to unlocking this potential lies in making sense of the internal data you have accumulated across multiple sources and linking all of them together.
As a matter of good practice, you would need to combine in one location, which is usually Data Lake, multiple sources of data, both structured and unstructured data, and create de-duplication rules in order to create a single source of truth. As much as possible, try to establish a connection, create a lineage from the data you have.
Some companies can be inundated, or even overwhelmed by unstructured data, but having that is still better than not having any. Besides, there are tools to help us make sense of the data, including the likes of Amazon, Microsoft, Google, and Alibaba, who all have data centers and offer services to help corporates to take advantage of their data.
Another factor in Singapore’s favor – the Government adopted a multi-pronged approach to cloud computing by leveraging on commercially-available public cloud capabilities. Keeping data in the Cloud enables companies to build new machine learning models at scale and faster speed.
Things to keep in mind when building AI culture in your firm
The financial industry is highly regulated for the right reasons, given the sensitivity of the data insurers and banks possess. The challenge for companies, therefore, lies in striking a balance between staying agile to take advantage of the opportunities presented by technology such as AI and making sure the confidentiality of our data is not compromised, particularly when we are working with large sets of data.
To move efficiently, I recommend companies to source for best-in-class AI solutions instead of trying to build all of them in-house.
To this end, it is good to have at least one group dedicated to understanding business pain-points across the company, and from there, find the most suitable AI-based solutions for these problems. Similarly, sufficient technical resources should be allocated to allow companies to prototype with these found solutions to figure out whether it makes sense to integrate them.
Indeed, if the intent of adopting AI or any other form of innovation is to solve business problems, then it only makes sense if the remit to innovate does not reside with just a particular department. Instead, every member of an organization should be encouraged to adopt an agile mindset, stay informed about the latest technology, continuously learn new tools, and be unafraid to take risks with new projects, even if there is a chance of failure.
If there is one thing I learned from working in different tech companies, it is finding ways to try something quickly and iteratively is often better than investing months or years to develop something that might or might not work.