RiskVert raises its Seed Round today and has closed the financing. The funding will go toward expanding our model engineering team, accelerating PAS integrations with Applied Epic and Duck Creek, and building out the ISO rating monitoring infrastructure that our pilot customers identified as their highest near-term priority.
This is an announcement, not a manifesto. The insurance industry gets enough of those. What I can tell you is why we started this company, what the product does, and what we are going to build with this capital.
Why We Started This
The problem we are solving started with a specific observation from our own experience working with regional P&C carriers. Underwriting decisions were being made with inconsistent data — not because the data did not exist, but because the tools to extract and analyze it were not part of the submission workflow. An underwriter reviewing an ACORD 125 for a commercial contractor was making a risk decision based on what was visible in the file, without knowing that the carrier's own loss history on that SIC code had deteriorated over the prior eight quarters, or that the territory the contractor operated in had been flagged as adverse in the most recent ISO loss cost filing.
That information existed. It was accessible to the actuarial team, buried in triangles and reserve analyses that are produced quarterly and distributed to senior management. It was not accessible to the underwriter at the point of the submission decision, and most carriers had no infrastructure to get it there. The result was pricing inconsistency, loss ratio drift that no one could fully explain, and a persistent gap between what the actuarial models projected and what the underwriting decisions actually produced.
We built RiskVert to close that gap. Not with a dashboard that underwriters ignore, but with risk scores delivered at submission intake, before the file is opened, that surface the relevant actuarial signals in a format that changes what the underwriter does next.
What the Seed Round Enables
The model engineering work is the core investment. Our submission scoring model is already in production with pilot carriers, and the early loss ratio tracking on scored versus unscored cohorts is favorable. The Seed Round funds the next phase of model development: improving the accuracy of the severity-weighted claims probability, expanding the training dataset to cover additional commercial lines beyond commercial property and CGL, and building the monitoring infrastructure that detects distribution shift between training-time data and production submissions.
The PAS integration work is where carriers have the most immediate questions. Applied Epic alone powers a substantial fraction of the regional carrier market, and our Epic integration has been our most common pilot entry point. Funding accelerates the development of the Duck Creek and Guidewire connectors, which expand our addressable market to carriers that are not on Epic but have the same underlying need for submission-level risk scoring.
The ISO rating monitoring capability — which flags submissions where the current manual rate diverges materially from the latest ISO advisory loss cost for the classification — is the feature that our pilot customers ask about most consistently. It is not a complex analytics problem; it is primarily an infrastructure problem of keeping ISO rate data current and making the comparison available in the underwriting workflow. That infrastructure is what we are building with part of this capital.
What We Are Not Building
Automated underwriting — systems that bind or decline without human review — is not on our near-term roadmap. Not because we do not think it is technically feasible, but because the carriers we are working with are not asking for it. What they are asking for is better information at the point where their underwriters make decisions, combined with portfolio monitoring that identifies emerging adverse selection trends before they develop into loss ratio problems.
The human underwriter with better information produces consistently better outcomes than either the unaided human or the automated system operating without human override. The evidence for this in other regulated industries — financial services, healthcare diagnostics — is consistent enough that we are comfortable building to support human decision-making rather than to replace it. Our pilot results support this approach: the carriers seeing the largest loss ratio improvement are the ones where underwriters are actively using the scores, not the ones where the scores are available but rarely consulted.
For Carriers Interested in the Platform
Our current pilot program runs for 30 days on historical submission data, without any production integration requirement for the evaluation period. We run our scoring model against your existing data, compare the model's predictions against actual loss outcomes on the same cohort, and present the backtesting results so you can assess accuracy before any deployment decision.
The onboarding process for production deployment typically takes six to ten weeks depending on your PAS and the scope of the integration. We support Applied Epic, Duck Creek, and Guidewire in production, and we maintain a documented API for custom PAS integrations.
If you are evaluating submission risk scoring or FNOL triage automation for your commercial P&C book, we would like to hear about your specific situation. Contact us at support@riskvertx.com and let us know which lines of business you are most focused on.
The Team
RiskVert is based in Simsbury, Connecticut. Our team includes actuarial professionals with FCAS and CPCU credentials, insurance technology engineers with experience at multiple major PAS vendors, and carrier operations professionals who have been on the receiving end of the problem we are solving. We are hiring for model engineering and customer success roles. Details are on the contact page.
Interested in a pilot?
Contact us at support@riskvertx.com or start a conversation.