
How Edelweiss Life Insurance is Redefining Industry Standards Through Innovation and AI
In this engaging interview with Research NXT, Abhishek Gupta, Chief Marketing Officer at Edelweiss Life Insurance, explores the remarkable evolution of the company, showcasing how they have harnessed the power of data and AI to transform its strategies across customer experience, risk management, distributor engagement, and personalisation. From overcoming initial scepticism to achieving successful AI implementation, Abhishek highlights the pivotal role of clear problem definition, patience, and iterative learning in driving impactful results.
Key Highlights from the Interview
- AI-Driven Innovation in Insurance: AI is transforming core functions such as fraud detection, claims processing, and customer retention. Predictive models help identify high-risk cases and customers likely to attrite, enabling timely interventions and improving persistency ratios.
- Hyper-Personalisation at Scale: AI empowers the creation of tailored insurance products that cater to specific consumer needs, making personalisation achievable at scale, which was previously unattainable.
- Challenges in AI Implementation: Overcoming scepticism, consolidating scattered data through projects like Dataverse, and ensuring timely action based on AI insights were key hurdles. A robust analytics committee ensures strategic prioritisation of AI use cases.
- Future of Insurance Marketing: Emerging trends include hyper-personalisation, proactive distributor engagement, enhanced fraud prevention, and improved customer communication through AI-driven automation and real-time insights.
“AI empowers us to make data-driven decisions, enhance operational efficiency, and ultimately deliver better outcomes for both the business and its customers.”
Technology has significantly evolved over the years, shifting the focus from automation to AI. With this transformation in mind, could you share how your role at Edelweiss has evolved and what key areas you are currently focusing on?
Abhishek: In essence, my role has not changed significantly in terms of external perception or designation. However, there have been notable shifts internally. Back in 2017, I was with Edelweiss Group, not directly with Edelweiss Life Insurance. Since then, the Edelweiss Group has undergone a major decentralisation, and now, each business operates fairly independently. As a result, I am currently part of the Edelweiss Life Insurance business.
In my role, I oversee marketing, customer experience, and training. However, my broader responsibilities go beyond these functions. Firstly, my primary focus is ensuring that we stay true to our purpose of protecting the dreams and aspirations of our customers. Secondly, an often-overlooked aspect of my role is making a difference in the lives of our distributors and employees, doing everything we can to support and empower them. Lastly, I work towards achieving the numerical commitments we have made to the board, ensuring that we meet and deliver on our goals.
While I don’t manage the top line directly, as part of the leadership team, I am actively involved in contributing to and supporting the overall commitments we have made as a business. This holistic approach defines my role within the organisation today.
In our previous conversation, you emphasised the importance of data enrichment and creating a unified view of the customer in your marketing strategies. How has your approach evolved since then?
Abhishek: Our approach is ever-evolving, driven by the continuously changing needs and behaviours of the outside world and our customers. Comparing the past to the present, we now have a much deeper understanding of our customers. We have comprehensive insights into their interactions with us, including formal and informal channels. This includes data on the number of policies they hold with us and even a clear view of their broader investment patterns.
While we’ve made significant progress, it’s also evident that the world is evolving at an incredibly rapid pace. What may have been our primary focus in the past has required adaptation to align with these changes. That said, we remain steadfast in our long-term vision: to protect the dreams and aspirations of our customers. While our methods and perspectives have evolved, our commitment to this purpose has remained constant.
Read More
With such a strong emphasis on understanding customer interactions and evolving needs, data-driven decision-making must play a significant role in your operations. How has this approach taken shape within your company?
Abhishek: Compared to earlier, many of our decisions now heavily rely on data, particularly its predictive insights. The availability of data, both within our ecosystem and from external sources like the Insurance Information Bureau (IIB), has significantly expanded. This shared data is instrumental, especially in areas like fraud prevention and risk mitigation.
The improvements in data availability have been matched by substantial enhancements in its authenticity. Our ability to predict consumer behaviour based on this data continues to improve daily. However, the current focus lies in translating these predictive insights into actionable strategies. It’s about tailoring actions, refining strategies, and aligning focus based on these predictions.
At the same time, we’re also evaluating the accuracy of our predictive models to ensure they align with real-world outcomes. This puts us in a phase of discovery—continually testing and refining these models to optimise their effectiveness and maximise their potential.
How is AI shaping your predictive models, audience segmentation, and targeting strategies?
Abhishek: AI is often associated with consumer segmentation or marketing, but its applications go far beyond that. It plays a critical role in backend processes, which indirectly impact customer trust and business outcomes. Let me illustrate this with two use cases.
Enhancing Claims Settlement Ratio:
One of the most important metrics for a life insurance company is the claims settlement ratio, which reflects the percentage of claims honoured. This number is vital because it showcases the trustworthiness of the insurer to prospective customers. Currently, our claim settlement ratio is at an impressive 99.23%, placing us among the top five in the industry.
AI plays a crucial role in maintaining and improving this ratio. By leveraging AI models, we can predict the likelihood of fraud in incoming cases during the risk assessment phase. These models analyse data points to flag cases that might pose a risk of fraudulent claims. This ensures that only high-quality business enters our portfolio, reducing the chances of rejecting claims due to undisclosed or misrepresented information. The outcome is a cleaner claims process, improved claims settlement ratio, and enhanced trust with customers, which directly impacts our ability to acquire new business.
- Reducing Customer Attrition:
In life insurance, customers often commit to paying premiums for 10-12 years, with average ticket sizes exceeding ₹1 lakh. Customer retention is critical for both profitability and ensuring the customer benefits fully from their policy. However, some customers may attrite before completing their payment terms.
AI helps us predict which customers are likely to attrite by analysing their behaviour and payment patterns. With this information, we proactively engage with these customers through regular follow-ups, reminders, and education about the benefits of their policy. Additionally, the sales team that onboarded these customers is alerted to maintain a closer relationship. This has significantly improved our persistency ratio, ensuring better financial outcomes for both the company and the customers.
These examples highlight how AI-driven models can address various business functions. While some applications, like fraud prevention, might seem backend-focused, they have a profound impact on the brand’s trustworthiness, profitability, and future customer acquisition. AI empowers us to make data-driven decisions, enhance operational efficiency, and ultimately deliver better outcomes for both the business and its customers.
How are you using AI to measure and track ROI across key areas like NPS and underwriting accuracy? Have you developed dashboards or models to assess its impact on business outcomes?
Abhishek: Currently, we focus on monitoring the behaviour and performance of our AI models, assessing how closely their predictions align with actual outcomes. Since these models are built on historical data, they require time and iterative improvements to enhance accuracy. For instance, a model that initially operates at 60% efficacy improves as more data is fed and adjustments are made, eventually achieving 80-90% accuracy. We consider a model performing at 85% or higher as optimal for our needs.
To oversee this process, we have an analytics committee responsible for evaluating and prioritising model development requests from various business teams. Since building a model is both time-intensive and costly, the committee conducts a thorough cost-benefit analysis before deciding which use cases to pursue.
Once a model is adopted, we commit to it for at least 1.5–2 years, allowing time for adjustments and improvements rather than abandoning it if initial results fall short of expectations. This approach ensures we refine models effectively and achieve the desired level of accuracy over time.
With the evolving governance and improved data sharing in the insurance industry, you’ve highlighted the enhanced quality and accuracy of information. How do you see the industry leveraging this data for both business and marketing use cases?
Abhishek: With the new Data Protection Act (DPA), we must be cautious about how we use data. While we’re still evaluating the implications of the Act, it is undoubtedly a step in the right direction, giving privacy back to consumers. As an industry, we need to evolve and create the best use cases around these new regulations.
On the availability of data, significant progress has been made, especially in risk management. Data sharing within the industry, such as through IIB, helps identify fraud, akin to how credit histories are shared via CIBIL. Additionally, customers willing to share their data enable us to offer more tailored solutions.
For example, a 25-year-old customer might purchase a term plan with a sum assured of ₹1 crore, locking in a low rate. However, as their income and responsibilities grow over the years, that initial coverage may no longer be adequate. Often, customers don’t revisit or upgrade their insurance, assuming the initial policy suffices. Through enriched data and advisor relationships, we aim to identify life-stage changes and offer upgraded or customised solutions to meet evolving needs.
While personalisation capabilities have improved, they still rely heavily on accurate and current customer data. We’re seeing gradual improvement in data availability and customer willingness to share, but there’s a long way to go. Currently, there’s still some guesswork involved in understanding a customer’s life stage and offering suitable products. Sharpening this process through better data will be key to enhancing personalisation and customer engagement.
Do you think the increasing prevalence of scripted and automated calls, particularly from insurance companies, has influenced how consumers perceive your brand or the insurance industry as a whole?
Abhishek: Our industry has faced a negative perception from the very beginning. Even before fraudulent calls became common, insurance advisors often struggled to get calls answered, and being known as an insurance agent frequently led to avoidance. This perception, while unfortunate, isn’t entirely unfounded. Historically, the industry has faced issues with mis-selling.
While most of the industry today operates with integrity, some cases of mis-selling persist—for example, offering savings products or term insurance to retired people, which may not be suitable. Term insurance, for instance, is primarily an income replacement product and isn’t relevant for those without an income.
That said, the situation has significantly improved. Mis-selling has reduced, as evidenced by the claims settlement ratios. Most life insurance companies now have settlement ratios above 85%, with private players averaging 90% or more. This reflects the cleaner data and quality of cases being handled. Although there’s still work to be done, I believe the industry is continuously improving.
What challenges have you faced while implementing AI for use cases like risk assessment, underwriting, and customer support? Were there any specific hurdles related to the industry, your organisation, or particular functions?
Abhishek: The biggest challenges we faced in developing and implementing AI models can be summarised as follows:
- Belief and Adoption: Convincing stakeholders to trust the process was a significant hurdle. Since the initiative was driven top-down, there was initial scepticism about whether the models would work. Early failures reinforced doubts, as both the team and the organisation were still learning how to effectively build and use AI models. Over time, consistent efforts and successful outcomes helped overcome this resistance.
- Data Availability and Consolidation: Data was scattered across multiple systems—customer acquisition, renewals, and redemptions were all stored separately. This fragmentation made it challenging to gather accurate, real-time data. To address this, we initiated a strategic project called “Dataverse,” aiming to consolidate all data into a single, unified source. It took three years, but we now have a reliable system with up-to-date, accurate data at our fingertips.
- Actionable Implementation: Once the models were developed, getting teams to act on the insights in real time proved challenging. For example, risk teams needed to prioritise flagged cases promptly for the feedback loop to improve model accuracy. Building a sense of urgency and aligning processes with the data outputs required time and effort.
- Prioritising Use Cases: With increasing demand for AI solutions across various departments, managing requests became difficult. It became crucial to clearly define use cases and prioritise those with the highest impact to ensure the models delivered meaningful results.
Overall, the journey involved addressing scepticism, building a robust data foundation, fostering timely action, and strategically focusing on the most valuable applications of AI.
What emerging trends do you foresee in insurance marketing over the next 3-5 years?
Abhishek: Personalisation has evolved significantly, moving towards hyper-personalisation. Historically, the industry operated on a one-size-fits-all approach, with the same product available to all customers. Now, AI enables us to customise products based on individual consumer requirements at scale, something that would be impossible without it. Here are the key areas where AI is making a significant impact:
- Hyper-Personalisation: AI allows us to design tailored products that meet specific consumer needs, driving greater relevance and satisfaction. Personalisation at this level was unthinkable previously due to scalability constraints.
- Distributor Engagement: Retaining distributors is as crucial as retaining customers, especially in an industry where over 85% of insurance purchases involve an intermediary such as a bank RM, life insurance agent, or financial services advisor. Given the inherent complexity and trust gap in insurance products, distributors play a pivotal role in the sales process. AI models are being developed to predict distributor attrition, enabling proactive interventions. While still in the refinement phase, these models aim to alert sales teams to potential distributor issues, allowing targeted efforts to address concerns and improve retention.
- Fraud Prevention: AI is enhancing fraud detection by identifying patterns and anomalies at the point of risk assessment, underwriting, and claims processing. This remains a critical use case, with continuous improvements to ensure accuracy and effectiveness.
- Customer Retention: AI-driven insights are helping us identify customers at risk of attrition, enabling timely actions to engage them, reinforce product benefits, and ensure their long-term commitment. This iterative process continues to improve persistence and profitability.
- Enhanced Communication: AI is transforming customer communication by enabling automated, personalised, and real-time interactions across multiple channels. This ensures that messaging is relevant and engaging, enhancing the overall customer experience.
These applications demonstrate how AI is reshaping the industry by driving efficiency, improving decision-making, and fostering stronger relationships with both customers and distributors. While some models are still being fine-tuned, the progress made so far is promising, and the potential for further innovation is vast.
What insights or industry benchmarks would you find most valuable from this report to help insurance leaders like yourself make informed decisions about AI readiness and technology adoption?
Abhishek: To approach AI implementation effectively, it’s essential to begin with the basics. Here are four key recommendations:
- Clearly Define the Problem: Take the time to stay in the problem and define it thoroughly before jumping to solutions. A well-defined problem leads to solutions that are closer to addressing the root cause. Avoid rushing into solutions without fully understanding the challenge.
- Be Patient: AI is a learning process—results won’t be immediate. Expect to experiment, iterate, and refine over time. Models will improve and become sharper, but patience is crucial during the initial stages of implementation.
- Champion AI Adoption: Be a flag bearer for AI within your organisation. Many people may be sceptical or misunderstand its potential. Advocate for its benefits and share examples of how it has delivered results in similar contexts.
- Embrace Experimentation: Success with AI often comes through trial and error. Be prepared to fail frequently, as this iterative process is critical to discovering solutions that yield meaningful, disproportionate results.