home / case studies / Asia Pacific property and casualty reinsurer (composite)
Insurance

Underwriting Workbench with Automated Submission Intake and Risk Enrichment

We built an underwriting workbench that ingests broker submissions, extracts key risk data and enriches it with external sources. Underwriters review prioritised, pre-populated submissions instead of re-keying emails and spreadsheets.

Asia Pacific property and casualty reinsurer (composite) · Insurance

Underwriting Workbench with Automated Submission Intake and Risk Enrichment

Challenge

Wildly inconsistent submission formats. We combined ML extraction with template learning per broker and a quick human correction loop that improved accuracy over time.

Approach

Broker emails and attachments (schedules of values, loss runs, slips) are parsed by a document extraction pipeline that normalises locations, values and loss history. Locations are geocoded and enriched with natural catastrophe hazard data, and a triage score ranks submissions by appetite fit and expected profitability. The workbench shows a single risk view, pricing model outputs, referral rules and authority limits, and records underwriting rationale for audit.

Outcome

Illustratively ~30–50% less time spent on submission data preparation Faster quote turnaround for brokers Consistent appetite and referral checks Better catastrophe exposure visibility at the point of underwriting Documented underwriting rationale for audit and portfolio review

~30–50%Illustratively less time spent on submission data preparation
✓Faster quote turnaround for brokers
✓Consistent appetite and referral checks

Problem solved

Underwriters spent much of their day on data preparation, response times to brokers were slow, and good risks were lost to faster competitors.

What we built

Broker emails and attachments (schedules of values, loss runs, slips) are parsed by a document extraction pipeline that normalises locations, values and loss history. Locations are geocoded and enriched with natural catastrophe hazard data, and a triage score ranks submissions by appetite fit and expected profitability. The workbench shows a single risk view, pricing model outputs, referral rules and authority limits, and records underwriting rationale for audit.

Benefits

Illustratively ~30–50% less time spent on submission data preparation Faster quote turnaround for brokers Consistent appetite and referral checks Better catastrophe exposure visibility at the point of underwriting Documented underwriting rationale for audit and portfolio review

Document extractionGeocodingCatastrophe dataTriage scoringUnderwriting workbench

Talk to us