AI & the Future of the Insurance Industry
No one ever accuses the insurance industry of being on the bleeding edge of technology, but our experts outlined a path forward that might help change that
At ITC Vegas 2025, EPAM’s thought-leaders participated in a panel discussion titled, “Harnessing AI for Strategic Innovation in Insurance.” The panel consisted of:
- Sandra Loughlin, Chief Learning Scientist, EPAM
- Steven Tesler, Principal, Insurance Consulting, EPAM
- Srinivas (JJ) Jonnada, CIO, RSL
- Kartik Sakthivel, Former CIO, LIMRA and LOMA
The conversation covered a wide range of topics from key trends and challenges in enterprise adoption to the importance of organizational leadership and culture alignment. What follows is a recap of some of the most profound and impactful insights from the thought leaders.
Legacy Systems Are Acting as an Anchor
The conversation opened with an exploration of the current state of AI in the insurance industry. Speakers were quick to point out that over the past six years, the industry has undergone significant modernization, especially in the realms of cloud transformation and data readiness.
However, as Sakthivel emphasized, there is still a lot of work that needs to be done to enable enterprises to become AI-native. “There are carriers in the US that have systems that predate the moon landing.” As he noted, those systems are not only hindering modernization efforts, but they’re also fueling the ongoing struggle within the industry around data quality as many organizations are left with fragmented, inconsistent data, limiting the effectiveness of AI transformation.
And for as much work that needs to be done on the technology front, there’s an equally sized workload that needs to be addressed on the customer experience front. Jonnada is quick to point out that insurance products are complex and often poorly understood. Enterprises should be aiming for a more a modern digital shopping experience for all stakeholders, a challenge all on the panel were in agreement will require a significant rethinking of current approaches.
Plenty of Pilots, Nothing to Show
From there, the panel began discussing key challenges and trends in the adoption of AI within the insurance industry. In particular, Loughlin made note of three key issues:
- There is currently a high degree of focus on AI in products and services, sometimes at the expense of other initiatives which could generate broader enterprise value.
- Many insurance companies are operating under the belief that merely training employees will be enough to encourage adoption, often ignoring myriad other factors that merit consideration in driving enablement.
- Too many executives are comparing themselves to current competitors and are failing to consider the capabilities and advantages of AI-native entrants.
Jonnada then then noted that these collective, short-sighted views play a role in why so many pilots fail to generate ROI. He then polled the audience, asking how many in attendance had advanced an AI initiative from the proof-of-concept stage to production, meaning it was generating tangible value. Unsurprisingly, no hands were raised. Jonnada’s advice was to focus on data, domain and processes as the key ingredients to generating value. As Tesler also noted, “Use-cases within the insurance industry are a mile wide and an inch deep.”
Change Management Will Only Become More Critical
The conversation then pivoted to the topic of change management. Of note was the notion that AI needs to be regarded as a catalyst for business process reengineering, rather than approached as just a tool for achieving a particular solution.
Loughlin was quick to make note that as much focus as the enterprise puts on technology, a similar level of care needs to be taken to address employee resistance to changing established roles and processes within the organization. As she pointed out, employees will likely resist sweeping change, especially if it’s implemented too quickly or clumsily, owing in part to simple human nature. She followed this up by noting success in AI adoption won’t just hinge on training; it will require the enterprise to rethink incentives, KPIs, organizational structure and culture.
Data as the Foundation of AI Strategy
As the conversation on AI evolved, the panel was in agreement that clean, accessible and well-orchestrated data is essential for generating value through AI.
However, as Loughlin pointed out, not all data lives on a server. “In every organization there are hundreds of data silos. But some of the most valuable data is locked in the heads of your people. When trying to understand and orchestrate change across your operations, try to access that data as well. Ask retirees, ‘How does our business work?’ Train small language models on company data and invest time and money in data orchestration and cleanup.”
The panel also noted that incumbents must be open to learning from AI-native competitors, especially as it concerns their unique approach to knowledge management.
Leveraging AI to Address Enterprise Headwinds
Tesler then asked the panel what advice they might give to CIOs and CEOs just beginning their AI journey. Sakthivel’s advice was to focus on quick wins and to prioritize only a handful of impactful use cases at a time. Loughlin built on this, adding that to help drive success in these use cases, leadership needs to involve a variety of stakeholders in the process, including IT, operations, data and people functions. Equally important, Srinivas added, the enterprise strategy around AI needs to be anchored in strategic outcomes and objectives.
Srinivas then elaborated, this holds true for both small and large firms alike, the difference being that small companies might want to start with more practical AI tools and education, whereas large firms can afford more experimentation. That said, he made particular note that large firms need to be more cognizant of not becoming bogged down in “pilot purgatory.” Laughlin then went on to note that regardless of the size of an enterprise, solving for data quality and people/culture issues were universal problems every organization will need to address.
The Importance of People
This theme of people and culture was a recurring touchstone during the remainder of the panel’s time together. Loughlin made an important point, noting that most organizations know more about their physical assets than they do about the skills and knowledge of their people. Her advice was for organizations to prioritize collecting data on employee skills and tasks to help guide AI investments and ensure adequate workforce planning, noting that a failure to address these issues could quickly derail the organizational culture.
The Roadmap to Value
In order for insurance companies to derive value from their investments in AI, they’ll need to be diligent about their approach. As is evident from the panel’s discussion, it’s not enough to simply invest in a particular AI model or platform and push it out to the enterprise.
There are a number of facets that need to be considered and planned for, including data quality, employee education and upskilling, change management and more. As Loughlin made note, enterprises should regularly be looking to their AI-native peers and market entrants to benchmark their progress.
For more information on EPAM’s insurance expertise, or to connect with EPAM’s insurance experts, click here.
