RateVault
A sandboxed workbook-to-service workflow that versions rating logic, compares deterministic outputs and requires actuarial and regulatory approval before use.
Carriers and managing general agents often maintain rating factors, commission rules and eligibility logic in spreadsheets. Exposing that logic through a governed service could reduce manual execution and preserve which model version produced an output. The supplied research confirms enterprise products that wrap spreadsheets or replace rating engines, leaving a possible mid-market layer.
Spreadsheet execution is not automatically deterministic or safe. Macros, external links, volatile functions, hidden sheets, iterative calculations and user-defined code may not translate. The supplied compiler candidate has an unresolved arbitrary-code-injection vulnerability, making untrusted workbook ingestion unsafe without strong isolation or a different engine. Even exact output parity does not authorize a rate: actuaries, underwriting, filings, contracts and regulators retain authority.
A workbook artifact, hash, source owner, formula graph, unsupported feature, compiled candidate, test fixture, parity result, actuarial finding, filing or regulatory approval, released version, quote request, calculated output, underwriting decision, policy issuance and claim outcome are separate. RateVault provides reproducibility, not actuarial or regulatory approval.
A carrier or managing general agent's rating, product or actuarial-operations leader maintaining business-critical workbook models but not replacing the full policy platform.
The product seeks to preserve workbook logic while any silent semantic difference can change a regulated price.
Insurance modernization and unbundled feature layers create a current opening.
Several related concepts support spreadsheet governance, though no direct connection is recorded.
The input identifies a costly insurance workflow, confirmed adjacent enterprise products and a plausible unbundled governance layer.
The buyer quartet is incomplete, workbook semantics are complex, a proposed compiler is unsafe for untrusted input, regulatory approval remains external and heavier incumbents exist.
Discussion
No comments yet — be the first to weigh in.
