Erasmus says;
- Data engineering was essential to enable his team to quantify claims instantaneously – and in order to effect that process SLVRCLD were fortunate enough to have access to millions of quotes on which to base their calculations.
- To automate as much as possible – something made essential by the pandemic – insurers must partner with specialist insurtechs.
- While many aspects of the claims process have improved, we are some way still from being able to catalogue a client’s house from video footage.
Why did you choose to solve pain points in the claims process specifically?
While on garden leave 19 years ago I was approached by a broker that was streamlining their claims payments with prepaid cards. While piloting we realised that insurers do not know what items to replace and which suppliers have stock and offer the best value.
We developed a tender solution that subsequently became the de facto content replacement platform in South Africa, because, unlike our competitors, we use supply and demand by grouping the purchasing power of all the insurers while making the processes as efficient as possible.
The reason we chose the name SLVRCLD is because we want to be the silver lining during tough times and also offer our solution through the cloud via API. It is also enjoyable knowing that you helped a claimant in their time of need.
How exactly has SLVRCLD leveraged technology to digitise the P&C claim process?
About four years ago, we realised that even though we have optimised the claims process dramatically, bringing it down from days to hours – and further to minutes for certain categories – insurers will need to start quantifying claims instantaneously as is expected by their always-connected clients.
This is only achievable through data engineering and fortunately we had access to millions of actual insurance quotes that we used as a basis to develop a large structured item dataset. With this we can, through various methods, accurately identify the claimed item, instantly determine the correct replacements and then quantify them at the insurer’s panel suppliers from quotes that were sourced beforehand. If no instantaneous pricing is available it breaks out into a quick tender process that is managed by our sister company, called Thesl, which is set up to handle large volumes.
Once the claim is quantified an insurer can then replace the items through virtual vouchers or filtered pre-paid cards ensuring that the items are brought back onto policy. If required, insurers can also make cash-in-lieu payments as that is sometimes the best settlement option.
What must the insurance industry do in order to succeed in the digital age, and how does SLVRCLD come into play?
During the lockdown insurers realised that there are many repetitive functions that are not suitable to work from home environments and are busy automating as many of these as possible. I believe this is only achievable by insurers partnering with specialist insurtechs that are set up to engage with insurers no matter where they currently are in their digitisation strategy. This will ensure that they can evolve with the insurer from manual, through augmented to robotic and end up at the holy grail of straight through claims processing wherever possible.
Further claim quantification is also moving closer to the client and performed at claim capture instead of assessing stage as has always been the case. This is through self-help portals where an insurer communicates with clients in their preferred channels while capturing the information in a more structured format. This speeds up claim finalisation by a factor of multiples, which creates trust and is the bastion upon which insurance is based. It further brings the claims processing costs down dramatically and should be a differentiating factor during the next five years after which it will become the norm.
Some innovative insurers are also incentivising their clients to create an inventory of large value items at policy inception to further speed up the claims process and prevent under insurance or upgrade fraud when claiming. This can also be used to communicate to clients more regularly if they need to adapt their policy and shows clients that you care.
In what other ways could machine learning be deployed to reduce friction in the P&C claims process, and where do you see the demand for such integrations to evolve in the insurance industry?
Machine learning is used to optimise, speed-up and personalise existing processes. This can be done through pre-filling data or capturing it in a structured usable format, eliminating repetitive functions or adapting the process around the client.
It is already used very effectively in claim replacement and repair quantifications, especially with visual assessments, chatbots and fraud analysis. I think it will also move up the value chain to pricing as risks could be determined more accurately in the future. Chat bots are also improving in leaps and bounds with more training data being generated and should become the preferred communication route in the future.
Unfortunately visual content item identification, outside of basic barcode and International Mobile Equipment Identity (IMEI) searches, are still quite a few years off for content claims but we are actively keeping an eye on it as we would love to be able to catalogue a client’s house from a video of them walking through their house.
Can empathy be built-into automated P&C claims processes, and where do you see the demand for such integrations to evolve in the insurance industry?
After an incident and recovery thereof one of the most important things for a claimant is to go back to normality with as little friction as possible. Therefore I think the best way to show empathy is to help a client capture, process and settle the claim as quickly as possible.
Insurers can be very innovative in their approach towards this such as done with parametric insurance where a client does not even have to lodge the claim. Further helping decrease the friction by capturing as much information in a structured format, with as little input from the client, to enable any laborious processes to be automated with the end goal of straight-through processing, where no human is involved in the process.
Insurers are starting to really prioritise client service and we are very fortunate to be participating at this inflection point to help improve their lives.
–
Christiaan Erasmus was interviewed by Kristoffer Lundberg, CEO at Insurtech Insights.
Join 160,000 members of the insurtech community by signing up for our newsletter here.
See our upcoming webinars here.