Reimagine Underwriting With Agentic AI at the CPCU Society’s In2Risk 2025

Artificial intelligence continues to change the way the insurance industry underwrites risk, manages claims, uses data, and communicates with policyholders. Agentic AI is the latest evolution of the technology, and it is being deployed in various ways to expedite the insurance process. A session at In2Risk 2025 will address the ways in which underwriting can be reimagined with Agentic AI.
Speakers Rajesh Narayan Iyer, VP of product and underwriting at ValueMomentum; David Kuhn, AVP and head of AI and emerging tech at ValueMomentum; and Elizabeth Gieseking, data science manager, model AI risk management at State Farm will take the stage to discuss this timely topic in their session, “Reimagine Underwriting with Agentic AI,” on November 5 at 8 a.m.
For more AI-related content at In2Risk, add these sessions to your agenda:
- The Human Edge of AI-Driven Insurance
- Revolutionizing Regulatory Compliance with Gen AI
- Leveraging AI and Analytics to Transform Underwriting and Claims Decisions
- AI in Action: Insights from Agents & Brokers on the Transformation of the Insurance Industry
- Leading with Heart in the Age of AI
Agentic AI and Underwriting Applications
David Kuhn spoke with us about applications of Agentic AI in underwriting, how it differs from AI-enabled tools already in use within the industry, and future possibilities for the industry once Agentic AI is widely adopted.
Risk & Insurance: How does Agentic AI differ from traditional AI or automation tools currently used in underwriting?
David Kuhn: Most of what we’ve called “AI in underwriting” has really been automation, rules engines, RPA, or predictive models that handle narrow tasks like filling in data or flagging anomalies. Agentic AI is different. It is not a point solution but a system that can reason across steps, decide how to achieve a goal, and adapt as context changes.
Instead of automating a single task, it orchestrates multiple steps in the workflow; retrieving data, applying guidelines, summarizing insights, and escalating edge cases, similar to how a junior underwriter supports the process.
R&I: Why is it important to distinguish Agentic AI from conventional automation projects? What’s the mindset shift required here and why?
DK: If we treat Agentic AI as just another automation, we risk underselling its potential and mismanaging how it is used. Automation projects are focused on efficiency, making something faster or cheaper.
Agentic AI requires a different mindset because you are giving AI systems autonomy to pursue objectives within clear guardrails. That means governance, oversight, and feedback loops must be built in, the same way you would supervise and train new underwriters. The shift is from automating tasks to partnering with systems that can adapt and learn.
R&I: How does Agentic AI improve underwriting quality versus just making it faster or cheaper?
DK: Speed and cost are easy wins, but the real value is in quality. Underwriting often suffers when data is incomplete or when large submissions are skimmed quickly. An agentic system does not tire or overlook details.
It can extract the right information, cross-check against guidelines, and surface exceptions consistently. That reduces leakage, improves risk selection, and strengthens compliance. The outcome is fewer misses, more consistency, and stronger decisions across the board.
R&I: Could Agentic AI play a role in judgment-based underwriting? How can it support or enhance human underwriting, rather than replace it?
DK: Yes. Judgment-based underwriting is where Agentic AI can add the most value as a support tool. It will not replace human judgment, but it can act like a skilled analyst preparing the file, bringing together external data, highlighting red flags, comparing similar risks, and running “what if” scenarios.
The underwriter still makes the decision, but with a clearer, more complete view of the risk. It is about empowerment, not replacement. We also utilize a concept of “adversarial collaboration” that allows agents to argue points against them. A human added to that argument allows critical thinking to win.
R&I: What does the future of underwriting look like if Agentic AI is widely adopted across the industry in the next 5 to 10 years?
DK: If adoption scales, underwriting will shift from a reactive, document-heavy process to one that is proactive and intelligence-driven. Underwriters will spend less time on manual review and more time on portfolio strategy, client engagement, and nuanced evaluation of risk.
Workflows will be continuously monitored by AI agents that flag changes in exposure or market conditions. The function will become faster, more precise, and more consultative, positioning carriers as risk advisors rather than just policy issuers.
R&I: If attendees walk away with one key takeaway from your session, what do you hope it is?
DK: That Agentic AI is not about replacing underwriters but about elevating them. With the right design and guardrails, these systems give underwriters more insight, more consistency, and more time to apply expertise where it matters most. The companies that embrace this partnership will set the pace for the future of underwriting.
Registration Now Open For In2Risk 2025
This year’s In2Risk event will take place in New Orleans from November 3-5 at the New Orleans Marriott on Canal Street. Register by October 17 to enjoy $200 off your registration with code SAVE200.