Insurance has always been a data-driven industry. Rapid technological upheavals in AI, automation, robotic-processing, machine learning, and the Internet of Things (IoT) have created an unprecedented volume of structured and unstructured data on constantly evolving consumer and customer demands.
Jeevan Gnanam, Founder and Group CEO of VeracityAI, discusses what this means for the insurance industry and its clients. VeracityAI has developed an AI platform called WENN CarEye that scans the exterior of a vehicle and analyses wear and tear, diagnoses problems and more. Gnanam is the immediate past chairman of SLASSCOM and has leadership roles in St. Anthony’s Industries Group, SAKS, HATCH and OrionStellar.
How can AI be used to address a particular business problem in the insurance industry?
One of the core preoccupations in general insurance is to understand how assets such as a building, a vehicle, or a piece of fixed machinery perform over a period of time, and how wear and tear can impact their valuations. A primitive yet effective way of doing this is to look at the asset to identify any wear and tear or damages.
We have taken this to another level by taking over two million vehicle images at various stages of deterioration or damage to train WENN CarEye to identify and understand with a high degree of probability what damages look like and their approximate estimates as well. We also train WENN to look at individual components of a vehicle which needed us to have a cost basis of each part so we can train WENN to understand the damage estimates for each of these as well. So, a lot of work goes into building an AI model to provide a solution as WENN does, and the application is not just for motor vehicles. AI can now understand 3D spaces from pictures, so it can understand if I want to insure my house. We have a quicker way to put the power of the AI application in the user’s hands and build trust in the technology by allowing them to take an image of a room in the house and gauge the number of square feet to understand how deterioration can take place over time. That is, one vertical of AI, which is computer vision, but many other things can also be done in terms of developing an AI model by using a limited partnership such as the one we have with WENN.
By using a limited partnership (LP), you can maybe automate some of the requests that come from the customer from mobiles or any kind of applications. Usually, 80% of customer inquiries can be handled by AI, and then LP becomes stronger to automate that part of the claims process as well.
There are several things to consider when developing an AI model but an important one that most technologists tend to underestimate is security and fraud prevention. Fraud detection is also critical in building an AI application and you have to think of all those aspects when you build a product.
Is AI a standalone solution? When an AI system keeps learning, because there is a constant feedback loop, especially about consumers and users, can that be fed into other processes as well, especially in insurance?
Our business model is very much on a per-claim basis, and the reason for this is the constant feedback when a new damage assessment comes in and we eventually re-train the model to take into consideration the new parameters. This model is very much a partnership approach, where we base our pricing per claim and there’s constant feedback with new images coming in and we retrain and re-finetune the AI model.
What are your experiences in encouraging insurance companies to take up WENN CarEye?
We looked closely at how AI can create value for major stakeholders in the insurance industry: insurance companies, their policyholders and third-party service providers such as garages. We helped insurance companies understand how AI can have a positive impact on profitability by reducing costs per claim. This is a quantitative benefit, and then there are qualitative benefits too such as enhancing customer experiences. Successful businesses always invest in delighting their customers, but in Sri Lanka, the insurance industry is mostly conventional and set in its ways. There is also a lack of transparency in areas such as internal procurement and costs for parts. For whatever reason, some stakeholders within the business do not want that kind of transparency which is a prerequisite for enabling the kind of AI solutions that we can provide.
There is a huge opportunity to build truly world-class AI products and solutions out of Sri Lanka: we have done this, and we like to continue doing so
How have you gone about building AI solutions with WENN and VeracityAI?
It has been a great partnership and journey with WENN and now we are taking WENN CarEye into Asia. When we started this journey, we had so many questions we could not have answered without a partner: for instance, we wanted to do automated damage claims, but we did not have the data to make that possible. All this had to be done using partnerships and working together, and I am very thankful for this. There is a huge opportunity to build truly world-class AI products and solutions out of Sri Lanka: we have done this, and we like to continue doing so.