Extract data from loss run reports
At renewal, before an underwriter quotes a commercial account, they ask for a loss run: the carrier-issued history of every claim filed against the policy, usually covering the last three to five years. A broker assembling a submission, an underwriter pricing the risk, and a risk manager tracking the account all read the same report, which pairs the policy header with a list of individual claims. Each claim carries a claim number, a date of loss, a description, a status, and the three money figures that decide the account's loss experience: the incurred amount, the amount paid to date, and the reserve still held open. A manufacturer renewing general liability and workers compensation (WC) coverage might present a loss run from a carrier such as The Hartford showing 14 claims over five years, a total incurred of $612,000 in USD, and two claims still open with reserves that could develop further.
The figures that carry the weight are the paid, reserve, and incurred split and the open-versus-closed status. Incurred equals paid plus the outstanding reserve: a WC claim opened 2025-04-17 with $52,000 paid and a $32,000 reserve carries $84,000 incurred, and that open reserve signals loss development that has not settled, which moves the price more than a closed claim of the same size. Claims group under a coverage type and a policy period, so a five-year run spans several policy numbers, each with its own effective and expiration dates and its per-occurrence and aggregate limits. A NAIC number identifies the carrier. A claim reported long after its date of loss flags late reporting that an underwriter reads carefully.
Talonic reads the loss run and returns the policy and carrier header plus a claims array, each claim with its number, date of loss, description, status, and its paid, reserve, and incurred amounts, grouped by coverage and policy period. A five-year run with 14 claims and $612,000 incurred loads into the underwriting workbook instead of a retyped ledger, so a broker can build the submission and the underwriter can price the renewal from structured data.
What gets extracted from loss run reports
How extraction works for loss run reports
Loss runs are issued by the claims systems at carriers such as The Hartford and Liberty Mutual, brokered through agency-management platforms such as Applied Epic, and forwarded as a scanned PDF read with OCR, so the claim grid and the placement of the paid and reserve columns differ by carrier. Talonic reads the report and maps it to the claims-history schema in the Field Registry, which separates the policy and carrier header from the per-claim detail. Each claim keeps its number, date of loss, reporting date, description, and status, and its paid, reserve, and incurred amounts are held as distinct figures so incurred can be reconciled to paid plus reserve. Claims are grouped under their coverage type and policy period, and each policy keeps its effective and expiration dates and its per-occurrence and aggregate limits. The carrier NAIC number is captured to identify the underwriter, and a deductible or self-insured retention is read where the run states it. Every value returns with a confidence score and pixel-region provenance under DIN SPEC 91491 conformity, so an underwriter can verify an open reserve against the source loss run before pricing the renewal.
Sample extraction
A five-year loss run covering two lines of coverage
{
"document_number": "LR-2026-0715",
"document_date": "2026-07-15",
"insured_name": "Cleveland Precision Machining LLC",
"insurer_name": "The Hartford",
"insurer_naic_number": "19682",
"currency": "USD",
"policy_coverages": [
{
"policy_number": "WC-2024-558102",
"coverage_type": "Workers Compensation",
"policy_effective_date": "2024-07-01",
"policy_expiration_date": "2025-07-01",
"each_occurrence_limit": 1000000
}
],
"claims": [
{
"claim_number": "CLM-2025-0417",
"claim_date": "2025-04-17",
"reporting_date": "2025-04-20",
"claim_description": "Lower-back strain, material handling",
"claim_status": "open",
"paid_amount": 52000,
"reserve_amount": 32000,
"claim_amount": 84000,
"coverage_type": "Workers Compensation"
}
]
}Frequently asked
Does it split paid, reserve, and incurred amounts?
Yes. The three are returned as distinct figures per claim, and incurred is reconciled to paid plus the outstanding reserve, because an open reserve is where future loss development hides and an underwriter prices it separately from what is already paid.
How are open and closed claims distinguished?
Each claim carries its status, so open claims and their reserves are surfaced separately from closed ones. An open claim reported long after its date of loss is flagged, since late reporting changes how an underwriter reads the account.
Does it group claims by policy period and coverage?
Yes. A five-year run typically spans several policy numbers, and each claim is grouped under its coverage type and the policy period it fell in, with each policy keeping its effective and expiration dates and its per-occurrence and aggregate limits.
What fields are extracted from loss run reports?
Talonic returns loss run reports as schema-validated, typed fields. Common fields include Report Number, Report Date, Named Insured, Insurer, and more, each normalized (dates to ISO 8601, amounts as numbers) and mapped to a stable key so the output shape stays the same across layouts.
How accurate is extraction from loss run reports, and how is confidence reported?
Every extracted cell carries a confidence score from 0.0 to 1.0 and a provenance pointer back to the source page and region, so low-confidence values can be reviewed against the original before the data is trusted downstream. There is no single accuracy number: confidence is reported per field so you can gate on it.
Can I use loss run reports extraction in production?
Yes. The same engine behind this guide is available as a production REST API and Node SDK with sync, async, and streaming modes, schema versioning, signed webhooks, and EU-resident processing. Start free with an API key, then scale on usage-based pricing.
What does it cost to extract data from loss run reports?
There is a free tier for prototyping and agent evaluation with no credit card. Paid usage is credit-based at 1,000 credits per euro: page ingestion is 100 credits per page and registry-resolved queries are free. See talonic.com/pricing for current rates.
Ready to extract from your own loss run reports?
- GDPR
- HIPAA
- ISO 27001 / 42001 aligned
- DIN SPEC 91491 co-author
- EU-resident infrastructure
Author note
Reviewed by Talonic engineering, schema review · last reviewed 2026-07-06