July 15, 2025
Speakers:
Stewart Reeder : Head of Insurance, EMEA, Hitachi Digital Services
Swati Trehan : Head of Strategy & Operations, Ema
Martin Donovan : Head of AI Solutions, Irish Life
The Core Concept of Agentic AI and Its Evolution
Agentic AI represents a significant advancement beyond traditional automation technologies. Unlike Robotic Process Automation (RPA), which operates within rigid, predefined workflows, agentic AI has the capacity to dynamically understand process flows and adapt its actions according to goals and evolving contexts. It brings a higher level of intelligence, enabling more nuanced and responsive automation. This technology can make decisions based on real-time inputs, provided appropriate governance and oversight are in place to ensure accountability. Agentic AI is designed to function in a human-like manner—capable of thinking, planning, and acting autonomously—without relying on centralized data. One of its key strengths lies in its ability to communicate effectively, not just with humans, but also with other AI agents, paving the way for collaborative, multi-agent ecosystems within enterprises.
Why Agentic AI is Critical for Insurers Now
The insurance sector, often hindered by data silos and legacy systems, is particularly well-suited for agentic AI. Earlier automation efforts in the industry were limited by inflexible tools and fragmented data environments. In contrast, agentic AI does not require centralized data and can seamlessly operate across various systems. Its rapid deployment capabilities mean that insurers can start realizing value in a matter of weeks rather than months or years. This quick time to value enables organizations to significantly boost productivity and scale operations without being constrained by workforce capacity or infrastructure. As the technology matures, those who adopt early are positioned to gain a competitive edge. On the other hand, delaying adoption could lead to greater integration challenges down the line, potentially making the transition more disruptive and costly.
Cutting Through the Hype: Reality vs. Overhyped Narratives
While agentic AI holds transformative potential, it’s important to differentiate between realistic capabilities and overinflated expectations. The technology has advanced rapidly and can deliver impressive outcomes, but it still requires a solid operational foundation to succeed. Data remains a critical enabler—it doesn’t have to be centralized, but it must be accessible and relevant to the use case. Additionally, agentic AI systems begin without any built-in context; they need to be trained through data, feedback, or a combination of both. A gradual, targeted approach works best. Organizations should initially focus on well-understood use cases with reliable data before attempting to solve more complex challenges. This methodical rollout helps build organizational trust and experience with the technology.
Practical Application and Use Cases in Insurance
The insurance industry is already exploring agentic AI in practical, business-driven ways. Implementations typically start with straightforward problems that offer measurable outcomes. For example, automating the handling of specific claims processes can demonstrate value in a low-risk environment. Improvements in speed and customer experience are immediate—processes that once took weeks can now be completed in a single day. Rather than replacing human workers, agentic AI enhances their efficiency, allowing employees to focus on higher-value tasks. Business users play a critical role in early implementation by contributing their domain knowledge to define relevant workflows. As confidence grows, organizations are expanding into more complex functions such as underwriting and pricing, proving that the technology is capable of evolving beyond basic customer service applications.
Overcoming Challenges: Data, Governance, Skills, and Organizational Change
Adopting agentic AI requires navigating several strategic challenges. The technology can work with data that is dispersed across multiple systems and can even learn from past behaviors to codify informal knowledge. However, it still depends on having some form of input—where no data or historical precedent exists, AI cannot operate. Governance and observability are critical for ensuring that decisions made by AI systems are explainable and auditable. Organizations are beginning to use AI tools to monitor and evaluate other AI systems, creating an additional layer of oversight. Centralized governance platforms are emerging to provide visibility and control across all AI initiatives, offering detailed insights into agent behavior and outcomes. As businesses move toward multi-agent environments, new protocols are being developed to enable secure and interoperable communication between agents, with attention to regulatory considerations such as GDPR. The rapid pace of AI innovation challenges traditional IT processes, requiring more flexible approaches to testing and deployment. In parallel, organizations must invest in identifying and empowering employees with the right mindset to drive AI initiatives. This includes rethinking organizational structures, training programs, incentive systems, and career paths to align with a future in which humans and AI agents collaborate closely. Everyone in the organization will need to become comfortable working alongside AI in a relatively short time frame.
Key Recommendations for Insurers
To capitalize on agentic AI, insurers should begin by identifying a clear, specific business problem to solve. A well-defined objective provides the focus necessary to guide both technical and strategic efforts. It is also essential to start now—early experimentation provides the practical experience needed to scale effectively. Collaboration between business leaders and technical teams is vital, with each side bringing its expertise to ensure the right problems are addressed in a scalable and governed way. Insurers should implement strong governance and observability frameworks from the beginning, particularly given the regulatory nature of the industry. Focusing on quick, tangible wins helps build trust and momentum, making it easier to justify broader adoption. Finally, preparing the organization culturally and structurally is key. Success with agentic AI hinges not just on technology, but on cultivating an adaptive mindset that embraces new ways of working, learning, and measuring success in a hybrid human-AI workforce.
Here’s a short AI-generated podcast for you to listen to: