Scout InsurTech Spotlight with Eric Fenton

Eric Fenton is the Principal Insurance Business Consultant at EPAM Systems, a leading global provider of digital engineering, cloud and AI-enabled transformation services and a leading business and experience consulting partner for global enterprises and ambitious startups. Eric was interviewed by Chris Luiz, CEO and Co-Founder at Scout InsurTech.

Eric, what are some of the less obvious or “non-obvious” ways AI is currently impacting the insurance industry?

“Much has been said about the pending industry talent gap of 400,000 vacant roles. I sounded the alarm in September of last year that with the start of the economic downturn and the influx of industry talent infusing into the insurtech landscape, carriers were faced with the reality of doing more with less.

Another focus area is AI Bias. As we know, Large Language Models (LLMs) are developed by a snapshot of the internet with enhancements on its ability to summarize by relative wording ranking. If we see bias in various ways through societal influence, it’s inherently included in historical data that populates LLMs. If we don’t test the positive and negative impacts through testing, we potentially proliferate bias at scale.”

How do you see AI helping (or hurting) the industry’s talent gap, particularly when many entry-level roles could be automated?

“Partnering with subject matter experts across EPAM, I’m focused on the end-to-end knowledge worker’s journey and its dependency on tacit knowledge. Addressing this elusive challenge head-on with AI has become a passion of mine. Gaining tacit knowledge out of someone’s head and digitizing with embedded workflow diagrams unlocks value across your enterprise. That’s the new gold rush in business consulting and organizational change management. 

As the industry gains less in fresh talent with limited knowledge, we must leverage technology to bridge that gap. It's imperative to work now to get ahead of the disruption and position the workforce for success. It’s important for organizations to understand how their AI strategy will impact employees and proactively create opportunities for upskilling, reskilling and internal mobility.”

.

Why is bias in AI models especially risky for insurers, and how can companies mitigate that risk?

“While no one intends to apply this with purpose, LLMs largely gain this bias from the collection of data they consume. As decisioning is applied through AI agents and AI, we must apply guardrails that apply responsible AI as a core principle. Being able to assess and apply correction is an area we are focused on. Decisioning is largely focused on underwriting, policy servicing and claims; these areas must be a priority for carriers. Leveraging forward-thinking vendors that have identified and corrected bias is key in mitigating this risk and applying the necessary strategy to avoid future incidents.”

What do you see as the greatest regulatory and compliance challenges facing insurers who adopt AI solutions?

“Since 2023, legislative language has been introduced within privacy laws, enabling regulators’ the ability to apply AI audits. If bias is applied as the baseline of any open source LLM model, how can we ensure bias is not applied to decisioning from underwriting to policy servicing and claims? Carriers must conduct an inventory assessment and effectively mitigate the risk of AI decisioning through deep analysis, adjustments and testing.

As the insurtech landscape has welcomed open-source LLM capabilities into their platforms as ‘features,’ carriers must keep an inventory of AI use and its impact on risk across the IT ecosystem. Additionally, the industry has to be vigilant about fraud (for example, fraudulent claimants are using AI to create false claims reports and even falsifying images to submit with the claim) and adverse selection.”

Looking toward the future, which AI-related developments should insurance leaders be keeping an eye on over the next few years?

“It may be hard to imagine, but we are still in the infancy phase of AI. If we look at the quick progression of LLMs to Agentic, we now see more options to develop code, apps and platforms using AI. Regardless of the application of AI to the business or IT Software Development Life Cycle, the landscape continues to evolve at an accelerated pace.

As quantum computing becomes more economical, security is top of mind. Quantum can crack the hardest security protocols we have today and globally; there are 100-200 online today. Focusing on fostering an AI-native organization, data and talent quickly comes to mind. Insurance leaders evaluate their ability to organize their enterprise data layer. This opens the door to AI-native products, operations and a deep understanding of their business.”

Scout InsurTech Thanks Its Presenting Partner

And Our Scout InsurTech Partners