Estimated reading time: 7 minutes
CFC’s Lane Assist moves real cyber submissions from email to quote recommendation in seconds. Every output is reviewed by a human underwriter before anything is issued.
CFC has launched Lane Assist, a live pilot of agentic underwriting within its cyber insurance team. The specialist insurer describes it as a world first in specialty insurance. The system takes a submission from email through to a quote recommendation in seconds. It is working on real new business submissions, in live workflows, producing real quotes today.
This is not a proof of concept. It is a controlled live deployment. That distinction matters for the broader market.
What Lane Assist Actually Does
Lane Assist automates two core tasks. It extracts data from incoming submissions. It then constructs a quote recommendation using CFC’s established underwriting rules and practices. The system targets low complexity, high volume submissions, the category that consumes underwriter time without requiring specialist judgment.
Every recommendation Lane Assist produces goes to a human underwriter for review, checking, and approval before anything reaches a broker. CFC has been explicit about that design choice. The pipeline accelerates the routine work. It does not remove human accountability from the output. The pilot launched this month with a small number of live submissions. CFC intends to test performance, learn from outcomes, and refine the technology before any broader rollout.
The Governance Model Is The Story
The speed claim will draw attention. Submission to quote recommendation in seconds is a striking number. The more important detail for underwriters, brokers, and risk managers is the governance architecture behind it.
Agentic AI systems act autonomously within defined parameters. They execute multi-step workflows without human input at each stage. That autonomy raises liability questions that the cyber insurance market is only beginning to address. Who bears responsibility when an agentic system extracts data incorrectly? What happens when a flawed quote recommendation clears human review and reaches a broker? CFC’s human-in-the-loop design is a direct response to those questions. It preserves underwriting judgment at the output stage while automating the upstream workflow.
Chris Mullan, Head of Data and AI at CFC, stated: “This is not a lab experiment — it’s working with real submissions, in live workflows.”
That framing is deliberate. CFC is positioning Lane Assist as a production system, not a research exercise. The implication for competitors is clear. Agentic underwriting has moved from theoretical to operational in specialty cyber insurance.
What This Means For Brokers
The practical effect for brokers is faster turnaround on straightforward risks. Lane Assist is designed to increase the proportion of submissions handled with minimal touch. CFC frames this as less friction, shorter response times, and greater confidence that routine risks are progressing efficiently.
The structural effect is more significant. If agentic underwriting handles routine submissions at scale, human underwriters shift their time toward complex risks and broker relationships. That changes the nature of the broker-underwriter conversation. Routine account management moves to the system. Specialist judgment concentrates on the cases where it adds the most value. For brokers placing complex cyber risks, that shift could mean better access to underwriter attention on the accounts that need it most.
CFC is also exploring how the time saved on routine submissions can be reinvested in broker relationships and complex risk assessment. That exploration is as consequential as the technology itself.
The Underwriting Implication Nobody Is Asking Yet
The Kroll 2026 State of Cyber Resiliency report, published today, finds that 76% of organizations experienced a security incident involving AI applications or models in the past two years. The report maps incident rates directly to cyber maturity. Low-maturity organizations experience AI incidents at an 89% rate. High maturity organizations hit them at 54%.
CFC is now running an agentic AI system inside its own underwriting operation. The cyber maturity of that internal operation, its governance framework, its human review controls, and its incident response capability is exactly the kind of variable the Kroll data says determines AI incident exposure. Lane Assist’s human-in-the-loop architecture reflects a high level of design maturity. But the broader question for the market is whether carriers deploying agentic underwriting will hold themselves to the same governance standards they require from policyholders.
That question is directly relevant to the affirmative AI coverage product category. If a carrier’s agentic underwriting system produces a flawed quote recommendation that clears human review, and that error leads to a coverage dispute or E&O claim, the liability chain runs through the AI system itself. The LLM agent research from MIT and Carnegie Mellon documented exactly this pattern systems executing tasks correctly by their own logic while producing outcomes their operators did not intend. Agentic underwriting introduces that risk class directly into the insurance production process.
Get The Cyber Insurance News Upload Delivered
Subscribe to our newsletter!
A Market Signal, Not Just A Product Launch
CFC holds a significant position in the cyber insurance market. Its decision to deploy agentic underwriting in production, on real submissions, sets a reference point for the market. Carriers watching this pilot will draw their own conclusions about timeline and risk appetite. Brokers will start asking other carriers when their own submission processes will follow.
Mullan added: “Our focus is on scaling agentic underwriting in a way that is genuinely valuable — for brokers, underwriters and ultimately our customers.”
The ransomware parallel applies here, too. When the first carriers began automating ransomware triage in the late 2010s, the rest of the market treated it as a niche experiment. Within three years, it was a baseline expectation. Agentic underwriting is earlier in that cycle. CFC’s live pilot suggests the cycle is already running. As AI risk expert Erin Kenneally warned in an earlier Cyber Insurance News podcast, the insurance market tends to move faster on AI adoption than on AI governance. Lane Assist’s design tries to close that gap. Whether the rest of the market follows that governance discipline as it deploys its own agentic systems is the open question.
FAQ: CFC Lane Assist And Agentic Underwriting
What is agentic underwriting?
Agentic underwriting uses autonomous AI systems to execute multi-step underwriting tasks without human input at each stage. In CFC’s Lane Assist pilot, the system extracts submission data and constructs a quote recommendation automatically. A human underwriter reviews and approves every output before it reaches a broker.
What does CFC’s Lane Assist actually do?
Lane Assist takes an incoming cyber insurance submission from email, extracts the relevant data, and produces a quote recommendation using CFC’s established underwriting rules. It targets low complexity, high volume submissions. The system is live on real new business submissions within CFC’s cyber underwriting team.
Is a human involved in the Lane Assist process?
Yes. Every quote recommendation produced by Lane Assist is reviewed, checked, and approved by a human underwriter before anything is issued to a broker. CFC describes this as augmentation of underwriting judgment, not replacement of it.
Why does this matter for brokers?
Agentic underwriting reduces turnaround time on straightforward cyber submissions. It also frees human underwriters to concentrate on complex risks and broker relationships. Brokers placing routine risks should see faster responses. Those placing complex risks may get more focused underwriter attention.
What are the liability risks of agentic underwriting?
Agentic AI systems can produce outputs that are logically consistent but operationally incorrect. If a flawed quote recommendation passes human review and leads to a coverage dispute or errors and omissions claim, the liability chain runs through the AI system itself. Human-in-the-loop governance, as CFC has designed, reduces but does not eliminate that exposure.
Related Cyber Insurance Posts
- What Your Cyber Insurance Policy Misses – And What Attackers Find First – NEW PODCAST
- Kroll: Low Cyber Maturity Drives AI Incident Rates To 89% – And Most Firms Are In The Middle
- Top Strategies for Identity Verification in the Age of Deepfakes, Remote Work, and AI Threats
- The Small Business Cyber Insurance And Cyber Security Reality Check – NEW PODCAST
- Small Businesses Must Prepare To Defend – And Insure – Against Cyberattacks
