“We have been using two decades of data in different ways aligned, not only with the ways the industry commonly looks at data but also in some unique styles suited to our organization.” Ayan De, Chief Technology Officer at Exide Life Insurance said.
“The data we were collecting got funneled into different use-cases over the years. One was for creating analytics models for the propensity of the business and the other was to create customer segments. We also used this data to understand the market better and churn out products in line with market requirements,” he added.
But once Exide Life started looking at the data sources, the company soon realized that it was not getting all the data points from the right sources into the analytics models. Another discovery was the quality level of this data. Though the company has been fine-tuning the applied weights and parameters, it is an iterative process and the team keeps testing its assumptions in the real market till they find a sweet spot that works for them.
With the market changing and evolving constantly, Exide Life continues to work on these models to remain on top of their game.
“We started in the manner that everyone does, with primary sources being the transaction data that was coming and landing in our core application. As we continued on the journey, we kept on adding more and more sources and we now have a full-fledged data lake as well,” Ayan said.
Taking control of the Data
“Not all sources will provide structured data and hence, we have invested in an ‘intelligent’ data lake. For example, the conversation happening between my contact centre executive and the customer is not structured data. Thus, a full-fledged data lake will take care of not only my transactional data but also the demographic view in addition to this kind of unstructured nebulous data. On top of that, we are investing in modern pieces of AI which make sense and pick up the right information from all these conversations and transactions to help us create the right customer persona,” he further added.
Exide Life has been able to use nearly 90 percent of the data points that land in its systems today, which, as Ayan mentioned, more or less takes care of the major business requirements.
It is not easy to achieve the Holy Grail of 100 percent data utilization because as the business flourishes and expands, the requirements for different platforms and applications continue to evolve. But the Company is building a roadmap to reach there.
“To increase the data quality and its use, we are now working and building a modern data lake and expanding, upgrading the current one on cloud where possibly, we will increase the percentage use of the available data. We still have the next step to cover which is the use of data from social media. That is something which has not been explored extensively yet, but we plan to reach there soon. We have already started testing some samples and as the data gets enriched, we are using more advanced AI/ML features,” he mentioned.
Fighting the Skillset problem
“Post lockdown, IT projects are being accelerated massively across the globe and all the tech companies have been trying to complete the already delayed projects for their clients at the earliest. Due to these phenomena, tech resources have become quite premium, and the market is seeing a major challenge in both resource acquisition as well as retention.” Ayan said.
He further said this has created a critical vacuum across industry verticals. Every industry today is having a problem with ‘in-sourcing’ tech resources specializing in modern technology stacks like Cloud, AI/ML, Analytics, and Middleware.
“But yes, we have run attractive programs in-house for learning and development plus active recognition of talent, helping build new tech capabilities and also retaining them as they grow in their roles and build aspiring careers with Exide Life,” he concluded.