Commercial GL declarations to schema-scoped JSON

Public Document Proof

A Commercial General Liability declaration page is turned into policy identifiers, dates, carrier and insured identity, limits, evidence, unresolved fields, guardrail metadata, and export-ready artifacts.

This is a product proof story based on public-document and benchmark behavior. It is not a customer deployment claim.

Status	Public-document proof case
Audience	Broker operations, MGA intake, compliance teams, and AI agents
Schema or surface	`commercial_gl`

All Case Studies Run Live Proof ->

Before / After

The job PolDex takes off the workflow.

The product boundary is narrow: turn messy insurance evidence or parser output into reliable, schema-controlled data that another system can use.

Before

Manual or generic extraction

Operators manually inspect declaration pages for policy number, period, carrier, named insured, producer, and liability limits.
Evidence is often copied into spreadsheets without page context or conflict state.
Downstream systems cannot tell whether missing fields are absent, unsupported, or low-confidence.

After

PolDex output

PolDex returns Commercial GL facts only, without unrelated personal-auto, homeowners, life, or other schema noise.
FastScript attaches page/source evidence to supported values and leaves weak values unresolved instead of borrowing evidence.
The result can be delivered as JSON, CSV, XLSX, webhook payload, processor review record, CLI output, or agent tool response.

Workflow

How the proof path runs.

Select the schema

The caller chooses commercial_gl through the API, processor, playground, live proof, CLI, or MCP.

Extract candidates

Readers collect document text and candidate fields while FastScript keeps the requested schema boundary active.

Adjudicate

FastScript normalizes values, checks evidence, strips unrelated fields, calculates guardrail state, and records unresolved critical fields.

Deliver artifacts

Customers receive structured JSON plus export links and evidence metadata that can be reviewed by humans or agents.

Output

What downstream systems can use.

These are the fields or artifacts this case study is meant to make legible to databases, workflows, operators, and AI agents.

Structured output

Data contract

policy_number
effective_date
expiration_date
carrier_name
named_insured_name
producer_name
each_occurrence_limit
general_aggregate_limit
products_completed_operations_limit
evidence
conflicts
unresolved_items
quality_guardrail_passed

Proof

Evidence contract

Source URL or upload provenance
Page or section reference
Visible evidence excerpt for each supported gold label
Required-field and evidence scoring through the benchmark path
FastScript abstention when a value cannot be defended

Try It