Why is it that in an environment so dependent on the richness of data, the insurance industry still relies on systems that are built simply to transact? Too many insurers claim to deploy modern, cutting edge technology solutions yet their systems are built to process actions without using the valuable insights from their data that can improve the user experience.
By building a system that only gets you from point A to point B, existing insurance systems become processing engines that don’t think. The guesswork is left to their staff and clients using the system, who often make their own subjective decisions as the system provides no insight, despite the volumes of data available to it.
These “dumb” systems are no longer going to cut it with the introduction of new InsurTech companies that focus on the customer experience and bring innovations throughout the insurance process, from sales to administration to claims.
The question that needs to be asked at the start of a system design is whether those setting up the processes are equipped with the skills needed to ensure that the system “learns” from the customer journey and is capable of using this to improve the process afterwards. In most cases they are not! As a result, the process is not as efficient and effective as it should be.
When setting up these systems, a clear differentiation needs to be made between the “actions” part of the machine and the “brains”. There is a difference between just “doing” and “processing, learning and improving”.
The challenge faced by decision makers tasked with implementing machine learning in their processes is identifying developers who can do two things – build a process and enable it to evolve. There are many smart developers who are great at creating task and goal orientated processes, but developers that create systems to analyse, observe and optimise have a completely different skill set – this requires a fusion of data science and technology.
For optimum productivity, data will be fed into a decision engine, models will be developed, algorithms will run, whilst simultaneously feeding back fresh insight into the system, suggesting real time changes to the user experience such as new product offerings, personalised incentives, useful marketing messages and improved claims handling.
If systems lack inbuilt data science capabilities, insurers will need to bring in external resources to analyse the data, often using (delayed) offline processes and with higher human costs and biases.
This can compromise the customer journey and reduce efficiency. Sales will be lost due to inappropriate focus, claims will be mismanaged due to routing delays and lapse rates increase due to inefficient retention strategies. The actual cost to the insurer ends up being much higher.
Decision makers that recognise the need to design systems to learn will shorten the insight loop by appreciating the value that real time business intelligence can add. They benefit from thinking systems as opposed to processing systems.
To be competitive, insurers should be asking how smart their systems are, and whether their processes learn and improve over time. If the answer is no, the guessing game will continue.