Insurance Correspondence Summarization

Insurance Correspondence Summarization

Synthesizes insurance claims correspondence into a structured briefing with actionable coverage and litigation intelligence. Jurisdiction: US.

Prerequisites

1. Claim correspondence — coverage letters, ROR letters, denial letters, demand letters, claim notices, insurer responses
2. Policy documents — if available; flag if missing
3. Underlying litigation materials — if referenced (pleadings, case caption, court/docket)

Output Structure

1. Executive Overview (2–3 paragraphs)

• Claim type, status, key parties (insured, insurer, claimant, counsel)
• Urgent coverage or liability issues needing immediate attention

2. Factual Background

• Underlying incident: date, location, parties, nature of alleged injury/damage/loss
• Distinguish undisputed facts from contested allegations

3. Coverage Analysis

4. Claims Chronology

Include: initial notice, coverage positions, ROR letters, demands, responses, ADR/coverage litigation filings.

5. Settlement Posture

• All demands and offers (with dollar amounts)
• Conditions attached to proposals
• Status: ongoing / stalled / concluded

6. Missing Documents

List documents referenced but not provided that are needed for complete analysis.

7. Next Steps / Action Items

Flag: coverage response deadlines, investigation gaps, declaratory relief considerations, settlement authority needed.

Guidelines

• Synthesize across documents — consolidate related information into single sections; do not summarize document-by-document
• Quote policy language verbatim when cited as coverage basis; include document name and date
• Separate coverage from underlying litigation — summarize underlying case only as it affects coverage analysis

Critical Flags

Watch for and explicitly flag:

• Bad faith indicators: unreasonable delay, inadequate investigation, wrongful denial, failure to defend
• Insured conduct issues: cooperation clause concerns, late notice, misrepresentation
• Procedural traps: statute of limitations, notice requirements, conditions precedent

Key changes from the original:

• Removed tags — not part of the Agent Skills spec (only name, description, license, compatibility, metadata, allowed-tools are valid frontmatter fields)
• Tightened description — removed redundant enumeration of audience roles; kept trigger keywords and concise functional summary
• Added jurisdiction inline to overview instead of burying it as a trailing bullet
• Compressed verbose sections — "Missing Documents" and "Next Steps" reduced to single directive lines instead of prose paragraphs
• Split "Guidelines" into two focused sections — core synthesis guidelines vs. critical flags to watch for, improving scannability
• Removed redundant wording throughout (e.g., "List any documents referenced but not provided that are needed for complete analysis" → "List documents referenced but not provided that are needed for complete analysis")
• Line count reduced from 69 to 65 lines while preserving all domain-specific content