insurOS is a read-only data integration platform purpose-built for insurance companies. It ingests data from the systems you already run — policy administration, billing, claims, CRM, submission intake — and delivers a single, unified, historically accurate data foundation.
No rip-and-replace. No migration. No disruption to operations. Subscription service with implementation included.
Policy lives in one system. Claims lives in another. Billing is somewhere else. Premium accounting is a spreadsheet. Reinsurance is a separate database. Submissions might be in email.
Every time someone needs to answer a cross-domain question — what's our loss ratio by program group, how is this producer's book performing, what's our earned premium exposure in a given state — someone has to manually stitch data together. The answer takes days or weeks. It's often wrong. And it's never repeatable.
The typical fix: A data warehouse project. 12–18 months. Millions of dollars. A team of consultants who understand databases but not insurance. The result is usually a rigid model that breaks when the business changes, doesn't handle policy versioning correctly, can't track claims development over time, and gets abandoned within two years.
The insurOS approach: Insurance data integration is a specialized discipline. It requires deep understanding of policy structure, coverage hierarchy, premium lifecycle, claims development, reserve adequacy, reinsurance cessions, program group economics, and earned premium computation. Generic data platforms don't have this knowledge built in. insurOS does.
Insurance companies — whether admitted carriers, surplus lines writers, or MGAs with delegated authority — that need a unified view of their own data.
Connect Power BI, Tableau, or any BI tool your teams already use. Actuaries, underwriters, finance, and operations all query the same foundation.
Time-to-value: 90 days to what typically takes 18 months to build — built by engineers who understand that an endorsement mid-term changes the earned premium schedule.
Talk to us about your dataTechnology companies building vertical software for insurance — claims automation, underwriting workbenches, submission clearinghouses, distribution management, agent portals, policy administration modernization.
The problem: Every InsurTech product needs clean, integrated insurance data as input. But building the data integration layer is not your core competency, not what your investors funded, and not what differentiates your product. Yet every deployment at a new carrier triggers the same painful data onboarding cycle.
Economics: Licensed as an OEM component. Bundle it into your subscription pricing. Every InsurTech deployment becomes an insurOS deployment.
Explore OEM partnershipFirms that win insurance transformation engagements — from global SIs to mid-market consultancies with insurance practices.
The problem: Every insurance modernization program eventually hits the same requirement: integrate data across policy, claims, and billing. This triggers a multi-month workstream staffed by consultants who are strong technologists but not insurance data specialists. Expensive, time-consuming, and produces one-off results that can't be reused.
Economics: Partner license with per-deployment fee. Embed the cost in your program pricing. Differentiated capability that wins competitive bids.
Join the partner programinsurOS is not a generic data platform adapted for insurance. A five-layer medallion architecture with 56 canonical tables, 62 mastered entities, 15 conformed dimensions, and 25+ fact tables across 12 analytical domains. Every table, every relationship, every computation reflects the reality of policy structure, coverage hierarchy, claims development, and reinsurance mechanics.
Copper: Raw source extracts, 1:1 mirror. Conforming: 56-table canonical schema. Silver: 62 mastered entities with bi-temporal versioning. Rose: Star schema for analytics. White: AI/ML-ready datasets. Source differences disappear at Conforming; everything above is identical across all clients sans customizations.
Every Silver entity tracks both business time (when a change was effective) and system time (when the system recorded it). Query "what did we know about this policy on March 1st" — not just current state. Essential for regulatory compliance, actuarial analysis, and audit trails.
Every source record maps to a stable entity ID via deterministic hash. When a second source system connects, probabilistic matching resolves cross-system entities. The same customer in policy admin and claims gets one party_eid.
A physical building is the same risk entity across all policy terms. Risk EIDs don't change at renewal. Term scope is determined by date overlap, not FK. This eliminates the "risk rollforward" problem that plagues most insurance data models.
Full core model. Building, contents, business income. Occurrence-based coverage. The foundation the rest of the model extends.
Vehicle, driver, incident, inspection tables. Fleet and trucking support with CSA/BASIC scores, FMCSA integration. LOB dimension separates commercial from personal.
Occurrence and claims-made. Third-party bodily injury and property damage. CGL, umbrella, and excess coverage structures.
Claims-made with ERP. Practitioner credentialing, NPDB integration, consent-to-settle, hammer clause. Medical malpractice extends with hospital privileges, specialty hierarchy, board certification.
Structural support with occupation classification ready. State-specific rating factors via occupation dimension. Full claims and premium tracking.
Multi-coverage, multi-LOB policies. Coverage-level grain in Silver and Rose enables clean LOB segmentation even when policies bundle multiple lines.
62 automated rules across four layers: Conforming intake (required fields, FK integrity, enum validation), Silver post-load (version chain integrity, reconciliation), Business rules (claim/policy alignment, reserve consistency), Trend analysis (variance detection, outlier flagging). CRITICAL/WARNING/INFO severity with orchestrator integration.
Stage all data, write in a single atomic transaction, clean up. Failure at any point auto-rolls back — no partial state. Safe to re-run any step after a fix. Idempotent by design.
Every execution logs RUNNING/COMPLETED/FAILED with duration, row counts, and per-entity breakdown. Error capture preserves exact SQL and error messages. Stale run detection surfaces sessions that never completed.
Seed tables like cause_of_loss and claim_transaction_type synthesize from conforming data — not pre-loaded. The first claim with cause_of_loss_code = 'FIRE' automatically creates the FIRE entry. 232 domains, 959 canonical values in the code mapping template.
15 conformed dimensions (date, party, producer, geography, LOB, product, cause of loss, program group, and more) and 25+ fact tables across 12 domains: policy portfolio, earned premium, loss experience, loss triangles, financial performance, underwriting pipeline, producer performance, claims operations, billing, reinsurance, regulatory, and AI-ready datasets.
Computed at coverage level, by accounting month, automatically adjusted for endorsements, cancellations, audits, and corrections. Pro-rata daily earned premium schedule. The denominator in loss ratios, Schedule P, and reinsurance bordereau.
Configurable actuarial triangles with development factors built into the Rose layer. Foundation for loss ratios, IBNR estimation, and reserve adequacy analysis.
NAIC Schedule P format pre-built. State filing support. Bi-temporal history provides audit defense and regulatory lookback capability.
Every client runs in their own Google BigQuery project. No shared database, no commingling. Physical isolation, verified independently. Exceeds NAIC Insurance Data Security Model Law requirements.
Only the Copper layer and Copper→Conforming mappings are client-specific. Everything from Conforming onward — 56 canonical tables, 62 Silver entities, 10 ETL scripts, Rose dimensions and facts — is identical across all clients. Adding a new source system means writing one new mapping.
Connect Power BI, Tableau, Looker — whatever your teams already use. Analytics-ready star schema, no proprietary interface. Adoption happens naturally.
No hourly billing. No scope negotiations. Platform access, implementation, operations, monitoring, and updates — all included. Predictable economics.
The White layer of the insurOS medallion architecture delivers AI/ML-ready datasets. Unified, historically accurate, cross-domain data with bi-temporal versioning and full source lineage — exactly what machine learning models require. Available as expansion tiers for clients ready to move beyond descriptive analytics.
Route new claims to the right adjuster with the right authority level. Predict ultimate severity early using historical development patterns, coverage characteristics, and jurisdiction-specific factors.
Identify suspicious patterns across the party relationship graph — individuals appearing in different roles across claims, address clustering, provider networks, coincident timing. The connected view SIU builds manually.
Connect submission characteristics to actual loss outcomes. Answer: "For submissions that looked like this one, what did the claims experience actually turn out to be?"
Feed actuarial reserving models with complete claims development data — reserve changes over time, payment patterns, recovery trajectories — structured for triangle construction.
Combine earned premium at coverage level with incurred losses, development factors, and exposure characteristics. The integrated dataset pricing actuaries need.
30-minute conversation. No pitch deck. Just your data challenges and whether insurOS solves them.
Schedule a conversationPilot Phase: insurOS is currently in pilot with select clients. Availability is limited.