In a sector grappling with over $100 billion in annual losses and the persistent challenge of legacy systems, the promise of digital transformation has often felt just out of reach. While rich in data, the insurance industry has seen only a fraction of its AI initiatives scale successfully. Here to demystify a path forward is Dominic Jainy, an IT professional with deep expertise in applying artificial intelligence and machine learning to solve these very challenges. Today, we’ll explore how a new class of technology, agentic AI, is uniquely positioned to bypass these historical roadblocks. We’ll discuss how it can transform the entire claims process from a frustrating series of handoffs into a seamless “resolve, not route” experience, the practical first steps for leaders to achieve measurable results, and the critical role of organizational structure in ensuring these powerful tools deliver on their potential.
With insurers facing significant annual losses and only seven percent successfully scaling AI initiatives, how does agentic AI specifically overcome the legacy infrastructure and fragmented data barriers that have stalled previous digital transformation efforts?
It’s a critical question because that seven percent figure is a stark reminder of a widespread struggle. The beauty of agentic AI is that it doesn’t require a complete “rip and replace” of the systems insurers have relied on for decades. Instead of being a passive analytical tool that needs perfect, clean data fed into it, an intelligent agent can be embedded directly into existing workflows. It acts as a smart layer on top of the old architecture, navigating the very fragmentation that stalls other projects. Think of it as a skilled assistant who knows how to pull information from different, clunky systems and make sense of it all to complete a task. This bypasses the bottleneck, allowing companies to start generating value immediately while they figure out their long-term infrastructure strategy. The key performance indicator isn’t just about efficiency; it’s about resilience in a market where high-frequency losses have become a structural reality that minor tweaks can no longer fix.
A “resolve, not route” approach in customer support sounds transformative. Could you walk us through how an intelligent agent manages a complex claims process end-to-end, from the first notice of loss to final communication, and what impact this has on loss-adjustment expenses?
Absolutely. The “resolve, not route” philosophy is a fundamental shift from the traditional, often frustrating customer experience. Imagine a customer reporting a loss. Instead of a simple chatbot that collects a name and then puts them in a queue, an intelligent agent initiates and manages the entire journey. It starts by capturing that first notice of loss, but it doesn’t stop there. The agent autonomously cross-references the policy, identifies that a specific document is missing, and immediately sends a proactive request to the customer. Once the document is received, the agent can update the policy and billing systems in the background and then notify the customer of the next steps. It’s a continuous, managed process. This has a massive impact on loss-adjustment expenses because it compresses the timeline dramatically. Every day a claim sits open, it costs money. By automating these steps and keeping the process moving, we’re seeing insurers slash those operational costs while simultaneously keeping the customer informed and, frankly, much happier.
We’ve seen metrics like a 23-day reduction in liability assessment time. For a leader looking to achieve similar results, what are the first practical steps to identify high-volume, repeatable tasks for automation, and how do feedback loops help refine these models?
Achieving a 23-day reduction is an incredible outcome, and it begins with a very pragmatic approach. The first step for any leader is to resist the urge to boil the ocean. You need to look for the low-hanging fruit: the high-volume, repeatable tasks that consume a significant amount of your team’s time. Think about processes like initial claims routing, document verification, or simple status inquiries. These are perfect candidates for initial automation. Once you’ve identified a process, you deploy the agentic model and build in robust feedback loops from day one. This is crucial. The model handles a task, and the human expert who would have normally done it either confirms the action or corrects it. That correction is fed directly back into the model, making it smarter and more accurate with every single interaction. This iterative refinement is how you go from a basic tool to a highly sophisticated system that can achieve those impressive reductions in cycle time and a 30 percent improvement in routing accuracy.
Since most scaling challenges are organizational, not technical, how can executives effectively establish an ‘AI Center of Excellence’? Please describe how such a center provides the governance and expertise needed to prevent fragmented adoption and build a true culture of accountability.
This is perhaps the most important point because, as you noted, about 70 percent of scaling challenges are organizational. An ‘AI Center of Excellence’ (CoE) is the central nervous system for this transformation. It’s not just an IT project; it’s a strategic business function. A CoE brings together the necessary technical expertise, business domain knowledge, and governance oversight under one roof. Its role is to prevent the “Wild West” of fragmented adoption, where different departments buy different tools that don’t talk to each other, creating new silos. The CoE sets the standards, vets the technology, and aligns every AI initiative with specific, measurable business goals. More importantly, it champions a culture of accountability. It ensures that the metrics for success are clearly defined and that the organization understands this isn’t just about technology—it’s about changing how work gets done to drive tangible returns. Without that central guidance, even the best technology will fail to scale.
Given the talent shortages in specialized roles like underwriting, how does agentic AI augment these experts instead of replacing them? Could you share an anecdote where this human-AI collaboration led to a more efficient and accurate outcome in a hard-to-fill position?
The narrative of replacement is a common fear, but in practice, agentic AI is all about augmentation, especially in highly specialized roles like underwriting or actuarial analysis where talent is scarce. These roles are incredibly hard to fill, and the experts are often buried in administrative tasks. An intelligent agent acts as a force multiplier for them. For instance, in claims, we’ve seen this with the Sidekick Agent, which helped claims professionals improve their efficiency by more than 30 percent. Imagine an experienced underwriter who would typically spend hours sifting through data and reports to assess a complex commercial risk. An agent can do that initial heavy lifting in moments—synthesizing data, flagging anomalies, and presenting a concise summary with recommendations. This frees the underwriter to focus their invaluable expertise on the strategic decision, the nuance, and the judgment calls that only a human can make. It’s not about replacing the expert; it’s about elevating their capabilities so they can handle more complex cases with greater accuracy and speed.
What is your forecast for agentic AI in the insurance industry over the next five years?
Over the next five years, I believe agentic AI will move from a competitive advantage to a foundational necessity for survival and growth in the insurance industry. We’re past the point of pilots and proofs-of-concept; the technology has proven its ability to deliver substantial ROI. I foresee a rapid acceleration in adoption, especially with the rise of industry-specific platforms that offer prebuilt frameworks, making implementation faster and more accessible. We’ll see intelligent agents become deeply integrated not just in claims and customer service, but across the entire value chain—from underwriting and risk assessment to compliance and marketing. The insurers who will lead the next era of innovation won’t be the ones who simply adopt the technology, but those who successfully build the organizational culture and scalable frameworks around it. For them, agentic AI will be the key to unlocking true operational resilience in an increasingly challenging market.