EZRater: insurance quote comparison software that cracked cross-carrier quoting

At the heart of the engagement was a deceptively complex problem: how do you build a comparative rating platform — a "Kayak for insurance" — when every carrier speaks a different data language?

When a broker or underwriter needed premium indications from multiple carriers, the traditional process meant logging into each carrier portal individually, re-keying the same risk data repeatedly, and manually collating results. For a busy MGA office handling dozens of submissions a day, this was an enormous productivity drain and a significant source of human error.

The vision for EZRater was straightforward: a user enters just three data elements, and within 8 seconds receives premium indications from every linked carrier simultaneously. But executing that vision required solving a deep technical challenge — getting heterogeneous carrier systems to accept, process, and respond to a single unified data submission.

The industry standard for insurance data exchange is ACORD (Association for Cooperative Operations Research and Development) — a structured XML-based schema that defines how insurance data should be formatted and transmitted. The problem was that while carriers nominally supported ACORD, each one implemented it differently: varying field mappings, non- standard extensions, different versions of the schema, and entirely divergent API architectures ranging from modern REST endpoints to legacy SOAP services and flat-file uploads. Building a true aggregator meant not just sending ACORD data, but dynamically transforming it into whatever dialect each individual carrier actually expected — reliably, at speed, and at scale.

The eight-second figure is the whole product. Beyond roughly ten seconds a user stops waiting and opens a carrier portal instead, which is the behavior the platform exists to replace. But the platform controls none of the systems it depends on.

Sequential dispatch is arithmetically impossible. Forty carriers averaging even half a second each is twenty seconds. Submissions have to leave in parallel, which means the slowest carrier — not the average one — sets the user’s experience.

So the deadline has to belong to the platform, not the carriers. A legacy SOAP endpoint having a bad afternoon cannot be allowed to hold a broker’s screen hostage. That forces an uncomfortable product decision: return thirty-seven quotes at eight seconds, or make everyone wait for the last three. Partial results within the deadline are far more useful than complete results outside it, provided the interface is honest about which carriers have not answered yet rather than quietly presenting an incomplete comparison as a complete one.

Failure has to be distinguishable from decline. “This carrier did not respond” and “this carrier will not write this risk” look identical in a results list and mean entirely different things to an underwriter. One is worth retrying; the other is information about the risk itself.

This is also why the adapter framework and the eight-second target are the same problem rather than two. Each of the 40+ carriers has its own mapping profile precisely so that translation is configuration applied at dispatch rather than logic executed per request — and so that a slow or failing carrier is one isolated adapter, not a fault in a shared path that degrades every quote in flight.

Running behind EZRater was the Straight Through Processing System (STPS) — a back-office management platform purpose-built to track, audit, and manage the full lifecycle of every quote sent and received through the aggregator. Where EZRater was the front-stage experience for underwriters and brokers, STPS was the operational backbone: logging every outbound ACORD submission, capturing carrier responses, flagging failures, and providing administrators with a complete audit trail of quoting activity across the office.

STPS gave MGA back-office teams visibility they had never had before — a single dashboard showing every quote in flight, every response received, and every submission that required attention. It also handled the enrichment pipeline: pulling pre-fill data from external sources, applying over 11,000 data quality repair rules, and ensuring that every submission reaching a carrier was clean, complete, and correctly structured before it left the platform.

EZRater fundamentally changed what was possible for surplus lines MGA offices. What previously required multiple portal logins, manual re-keying, and 20–30 minutes of effort per submission was reduced to a three-field form and an 8-second response. The cross-carrier ACORD adapter framework — the hardest engineering problem in the project — became a durable competitive moat: as new carriers were onboarded, each integration compounded the platform's value without proportional engineering cost. Behind it all, STPS gave back-office teams the operational visibility and audit control they needed to manage high submission volumes with confidence.

Today, Pre-Fill Services' products are embedded in nearly 500 MGA offices across every US state. The PDFReader alone, in production for over 20 years, has processed countless ACORD submissions backed by a data quality repair engine that continues to grow. The broader suite represents what becomes possible when 40 years of insurance domain expertise is paired with engineering that attacks the hardest interoperability problems at the root.