How do you automate loss run processing?
Automating loss run processing means reading the carrier reports into a consistent claim-level summary that someone can verify. Start with the original files, identify the policy periods and valuation dates, extract the relevant claim details, and reconcile the output with the source totals. The useful result is a clearer submission or analysis, with less manual transcription.
| Step | Prepare | Review |
|---|---|---|
| Collect reports | Gather the relevant carrier reports and policy periods. | Confirm the account, date range, and whether the reports are current. |
| Extract claim details | Organize claim number, loss date, description, paid, reserve, and status when present. | Compare each value with its source row and heading. |
| Preserve context | Keep the carrier, coverage, policy period, and valuation date with the claims. | Do not combine unlike totals or treat different valuation dates as identical. |
| Reconcile | Compare the extracted rows and totals with the original report. | Investigate omitted rows, duplicate claims, negative adjustments, and unclear amounts. |
| Prepare the next output | Build the reviewed summary needed for the submission. | Have the appropriate person interpret the loss history. |
The longer version
Carrier loss runs vary in layout and terminology. A spreadsheet, a scanned report, and a PDF with several coverage sections can require different review. Test the documents your agency actually receives, including a difficult file, before planning a high-volume process.
Define the output before you extract it. A submission summary may need claim dates, descriptions, paid amounts, reserves, and open or closed status. Keep the original headings and valuation date so a reviewer can understand what each number represents. Confirm the years and report requirements with the intended market.
Relay can help read supplied loss run documents and prepare a summary in the task. Ask it to identify the source for each material value and flag information it cannot establish. This is document preparation, not a promise of a dedicated loss-run database, automatic claim ingestion, or a confidence score on every field.
Measure missing or incorrect rows, review time, and the effort to produce the final summary. A clean table is not enough if someone has to reconstruct it from the original report afterward.
Common questions
What information should we extract?
Choose the fields needed for your purpose. Common examples include claim reference, loss date, description, paid amount, reserve, status, policy period, and valuation date when present in the source.
Does extraction request the loss runs from the carrier?
Collecting the report is a separate step. This workflow starts once the agency has the relevant files.
Can Relay help with a loss run?
Bring the report into the task and ask for the specific summary you need. Check the extracted rows, terminology, and totals against the original before using it.
How do we judge whether an extraction tool is useful?
Run several of your own reports. Count missing or incorrect rows and the minutes needed to reconcile the output, including difficult scans and unusual formats.
Related questions
Source notes
Sources checked 2026-09-05. Product details can change.
Part of the Relay straight answers library. Updated 2026-07-11. See how we source content.
Your next step starts with one email.
Join the waitlist. We’re inviting agencies regularly and will email you when access is available.
Join the waitlist