Structure the unstructured.
Every underwriting decision is a data decision first.
The model gets validated. The data gets trusted.
Everyone scrutinizes the catastrophe model. Almost nobody scrutinizes the data feeding it. Mismatched columns, missing codes, a thousand formats for the same field, and 1 to 3 business days of cleanup while the quote goes stale.
Automation without verification is risk.
Straight-through processing is only as good as the data beneath it. Speed you cannot audit is not speed. It is exposure.
Four years of human eyes. Then automation.
Our analysts verified submissions line by line for four years before a single SOV flowed straight through. That history is what our AI learned from. Every submission is scored like a golf hole: on a Par 3, the AI takes a hole in one and the file goes straight through. Par 4s and 5s call in the experts.
Simple files, cleared in seconds
Standard schedules validated, mapped, and geocoded automatically. Every step visible.
Complex files, guided in minutes
Ambiguous fields are flagged, not guessed, then resolved with expert review you can see and override.
The hard ones, handled within an hour
Multi-tab monsters go to the analysts who trained the system in the first place.
Underwriters should underwrite. Not clean spreadsheets.
Underwriters judge risk. The cleaning happens on the Wizard engine, run by our analysts or licensed to yours. The submission arrives in Looking Glass already structured, already verified, every assumption on display. Full control, none of the labor.
One philosophy. Two products.
SOV Wizard
THE ENGINE · FOR DATA OPERATIONSThe engine our own analysts run every day, trained on four years of verified submissions. License it and structure accounts in-house. Onshore, for less than offshore.
Geocodes, rooftop footprints, and exposure distributions, straight from the SOV.
Missing fields filled from reference data. Exceptions surfaced for sign-off.
Every mapping and transformation visible before it hits the model.
Verified output flows to modeling while the submission is still warm.
APIS TO CATASTROPHE MODELS, AND OTHER SYSTEMS
Looking Glass
FOR UNDERWRITERSFor the first time, the portfolio tells the truth. Push a raw submission and you are inside its dashboards four seconds later.
Submissions arrive filtered by your underwriting appetite. What you were always going to decline never costs you an hour.
Account and portfolio exposure, live, before the risk is bound.
Every location on the map, full GIS built in, every metric one click deep.
Remodel the cleansed account directly under varying policy terms and watch the numbers move.
One hundred submissions. Three ways.
Sixty simple files. Twenty-five medium. Fifteen complex. The same book, run three ways, timed in production.
MEASURED PROCESSING TIMES · SIMPLE: 20 MIN MANUAL, 1 MIN WIZARD · MEDIUM: 1 HR MANUAL, 5 MIN WIZARD · COMPLEX: 4 HRS MANUAL, 10 MIN WIZARD
Automation shouldn't stop when complexity starts.
Transparent from submission to decision.
Data transformation
Raw submissions turned into structured, modeling-ready data you can trace.
Catastrophe modeling
Hands-on CAT modeling and import file prep, with assumptions you control.
Underwriting workflow automation
Automated intake, triage, and exposure review. The underwriter stays in control.
Rater modernization
Macro-based raters rebuilt as stable web tools. Submission data populates rating files instantly.
Data, meet discipline.
The model is never the first step. The data is.
Send us your ugliest SOV. If we cannot structure it, you will know exactly why. That is the point.