Podcast: how AI is reshaping specialty insurance underwriting.
In Episode 305 of the FNO: InsureTech Podcast, Andrew Yeoman, Co-Founder and CEO of Concirrus, joins hosts Rob Beller and Lee Boyd to discuss how AI is reshaping specialty insurance underwriting.
The conversation explores the shift from AI experimentation to real operational deployment, including how insurers are reducing manual workload, accelerating quote response times, and scaling underwriting operations without increasing headcount.
Andrew also shares Concirrus’ perspective on what comes next for the industry, from domain-specific AI and agentic workflows through to entirely new insurance operating models enabled by AI.
Key Topics Covered AI in Production, Not Just Proof-of-Concept Andrew discusses why many AI underwriting initiatives struggle to move beyond experimentation, and why operational deployment requires more than simply adding a large language model into existing workflows.
The discussion reinforces a key industry reality: AI alone is no longer the differentiator. The real challenge is orchestration, governance, explainability, and embedding AI into underwriting operations in a way underwriters can trust.
This closely aligns with Concirrus’ broader view on building production-grade AI systems for specialty insurance.
From 48 Hours to 90 Seconds
One of the standout moments in the episode is the discussion around quote responsiveness and underwriting speed.
Andrew explains how Concirrus has helped insurers reduce submission-to-quote processing times from 48 hours to approximately 90 seconds through AI-driven workflows and intelligent automation.
The episode also highlights how operational efficiency can directly unlock underwriting growth, including examples where insurers scaled premium volumes significantly without increasing headcount.
Why Domain-Specific AI Matters
A major theme throughout the podcast is the importance of domain expertise inside AI systems.
Andrew explains why generic AI models struggle with specialty insurance workflows, where terminology, document structures, and underwriting logic vary significantly across lines such as aviation, marine, construction, and crisis management.
This mirrors many of the technical challenges outlined in Concirrus’ research around submission automation, including:
- semantic ambiguity across insurance terminology
- contextual inference requirements
- orchestration across multiple AI models
- explainability and governance requirements
- handling large unstructured submissions at scale
The Rise of Agentic AI in Underwriting
The conversation also explores the emergence of agentic AI systems.
Rather than relying on a single model, Andrew discusses how underwriting workflows increasingly require multiple AI agents working together across extraction, validation, reasoning, consistency checking, and decision support.
This reflects the architectural direction Concirrus has been building toward internally through multi-agent underwriting workflows and orchestration layers.
AI-Native Is Becoming Table Stakes
Andrew offers a candid perspective on the market’s increasing use of the phrase “AI-native,” arguing that the label itself is rapidly losing meaning as AI becomes embedded into almost every software platform.
The more important differentiator, he suggests, is whether companies can operationalise AI successfully and create measurable business outcomes.
That includes:
- improving underwriting responsiveness
- reducing manual workload
- improving consistency across portfolios
- enabling underwriters to focus on higher-value decision-making
The Future of Insurance
The latter half of the episode looks further ahead at how AI may fundamentally reshape the insurance industry.
Topics include:
- the possibility of AI-enabled “one-person” MGAs
- adaptive and dynamic insurance products
- real-time risk intelligence
- the role of AI agents in underwriting operations
- why specialty insurance may become one of the most important proving grounds for enterprise AI
Andrew also introduces the idea of “but for” companies, businesses that only exist because AI makes them possible.
The insurance industry is moving beyond AI pilots and experimentation. Why This Matters
The market conversation is now shifting toward:
- operational deployment
- workflow integration
- governance and explainability
- measurable underwriting impact
- scalable production-grade systems
This podcast provides an inside look at how those challenges are being approached inside specialty insurance underwriting today and how underwriting automation is moving from experimentation into operational deployment.
