Simon Glairy is a distinguished authority in the evolution of commercial insurance, bringing over two decades of expertise in risk management and Insurtech strategy. As the industry faces a dual transformation—an unprecedented “silver tsunami” of retirements alongside the rapid integration of artificial intelligence—his insights provide a critical roadmap for maintaining technical excellence. Glairy’s perspective bridges the gap between the intuitive wisdom of veteran underwriters and the data-driven efficiencies of modern technology, offering a masterclass in strategic talent management and capital preservation.
Approximately 30% of experienced insurance practitioners are expected to retire within the next five years. What specific “unwritten” knowledge regarding market cycles is most at risk of disappearing? How can firms systematically capture these nuances before experts leave, and what happens to pricing discipline if this transfer fails?
The knowledge at greatest risk isn’t the kind found in a textbook; it’s the “muscle memory” of navigating prolonged soft markets and the sudden, jarring corrections that follow. Veterans have a deep, lived understanding of how specific wording adjustments truly alter exposure over a decade, or how pricing discipline tends to erode when competitive pressures peak. To capture this, firms must move beyond informal chats and create structured mentorship programs where senior leads walk juniors through “post-mortem” reviews of difficult renewal cycles and complex claims years. If this transfer fails, we risk a “memory-less” market where pricing becomes reactive and volatile, leading to severe technical leakage because the new generation hasn’t seen how today’s small concessions become tomorrow’s massive loss trends.
Market relationships often dictate how much flexibility an underwriter can responsibly apply to a complex placement. How do you distinguish between data-driven risk assessment and the “relationship capital” built over decades? What specific steps ensure this interpersonal insight remains part of a firm’s institutional memory during succession?
Data-driven assessment tells you the “what” of a risk, but relationship capital tells you the “how” and “why” behind the broker’s submission. It involves knowing a broker’s historical transparency, the client’s long-term risk appetite, and where flexibility can be applied to achieve a sound commercial outcome without compromising the firm’s integrity. To preserve this, succession planning must be integrated into the daily workflow, requiring senior underwriters to bring their successors into high-stakes negotiations and broker strategy sessions. We must document these qualitative insights—such as a broker’s tendency to highlight or hide certain loss drivers—within our internal systems so that the relationship is with the firm, not just a departing individual.
AI is increasingly used for submission triage and data enrichment to handle rising volumes. Where exactly should the line be drawn between automated efficiency and human accountability for capital allocation? How do you ensure that AI enhances a clean underwriting process rather than simply masking a messy one?
The line is drawn at the point of ultimate decision-making and capital commitment; AI is an exceptional navigator, but it is never the captain of the ship. While AI can standardize data from multiple formats and flag inconsistencies across thousands of submissions, it cannot assume the moral or financial responsibility for a segment’s aggregate exposure. To avoid masking a “messy” process, insurers must first streamline their submission handling and triage criteria manually before layering on automation. You cannot automate chaos into order; you must have a clean, logical workflow that the AI then accelerates, ensuring the technology reinforces consistent decision-making rather than just speeding up bad habits.
There is a concern that junior underwriters might become mere reviewers of system-generated outputs. How can firms structure training so new talent learns to analyze risks and defend independent decisions? What specific exercises or mentorship models can prove they are developing a nuanced, rather than narrow, perspective?
We face a real danger of “deskilling” the next generation if we only teach them to click “approve” on an AI’s recommendation. Training must be structured to require junior underwriters to articulate the technical reasoning behind their calls, using the AI-provided data—like prior loss experience and portfolio positioning—as evidence rather than an answer key. One effective exercise is the “blind review,” where a junior underwriter assesses a risk manually before seeing the AI’s triage output, followed by a defense of any discrepancies before a senior peer. This develops a “richer context” mindset, ensuring they understand the underlying risk drivers instead of just trusting a system-generated score.
Implementing new tools often adds layers of work rather than reducing the burden on staff. How should an insurer rethink their workflow before layering on automation? Can you walk through a scenario where AI-driven context—such as prior loss experience—successfully guided an underwriter’s final call on a difficult risk?
The most successful implementations start by identifying where underwriters spend time on “low-value” administrative tasks, such as re-keying data, and removing those friction points first. Imagine a scenario involving a complex property placement where the initial submission looks standard, but the AI instantly surfaces a decade of loss development data from a similar occupancy type that the broker didn’t highlight. The underwriter sees this context immediately, allowing them to shift their focus from data gathering to a high-level negotiation on sub-limits or exclusions. This transforms the AI from a burden into a strategic partner that provides the ammunition needed to make a difficult, yet technically sound, final call on pricing.
What is your forecast for the future of underwriting judgment?
I believe we are entering an era where underwriting judgment will become more critical, not less, as the “commoditized” aspects of the job are absorbed by technology. My forecast is that the most successful firms will be those that treat AI as a cognitive bridge, using it to accelerate the development of junior talent by giving them access to decades of institutional data in real-time. We will see a shift toward a “bionic” underwriting model where the human’s ability to navigate regulatory shifts and complex broker nuances is supercharged by machine-led data enrichment. Ultimately, AI will not replace the underwriter, but it will certainly replace the underwriter who refuses to use it, creating a more engaging and high-stakes professional environment for the next generation.
