Reinsurance Portfolio Acquisition Diligence

The Problem

Acquiring a legacy or run-off reinsurance portfolio means inheriting its documentation. That documentation is rarely clean.

A typical portfolio acquisition diligence package includes bordereaux from multiple cedants, actuarial reserve reports, policy wordings and endorsements, loss runs from third-party administrators (TPAs), claims correspondence, and regulatory filings. These documents arrive in different formats (PDF, Excel, scanned images), use inconsistent terminology, and span years or decades of underwriting history.

The diligence challenge is structural, not just volumetric. Three problems compound:

  1. Reserve adequacy assessment requires reconciling paid, incurred, and outstanding reserve figures across cedants and TPAs that report in different formats, currencies, and development periods. Manual reconciliation is slow, and discrepancies often surface late in the transaction cycle, after pricing assumptions are set.

  2. Pricing support depends on identifying severity patterns, loss development trends, and exposure concentrations buried across thousands of claims files and actuarial exhibits. Manual sampling cannot cover the full portfolio, and missed patterns lead to mispriced risk that erodes combined ratios post-acquisition.

  3. Regulatory reporting obligations attach immediately on closing. Acquirers must produce standardized reserve reports, solvency calculations, and scheme documentation from the same fragmented source materials. Delays in structuring this data create regulatory exposure.

Traditional approaches rely on teams of analysts manually cross-referencing spreadsheets, extracting figures from PDFs, and building reconciliation workbooks over weeks. This is expensive, inconsistent across analysts, and produces outputs that are difficult to audit or reuse in subsequent reporting periods.

How Parsewise Addresses It

Parsewise operates at the document-package level, ingesting entire portfolio diligence sets and reasoning across all documents simultaneously rather than processing them one at a time.

Heterogeneous document ingestion

The platform accepts the full range of formats found in reinsurance diligence packages: PDFs (text-based and scanned), Excel bordereaux and triangles, Word policy wordings, PowerPoint actuarial presentations, and images. All formats pass through the same extraction pipeline, preserving tables, figures, and structure. Parsewise supports over 70 languages, handling cross-border portfolios where documentation arrives in multiple languages across jurisdictions.

Loss run and triangle standardization

Parsewise automatically standardizes loss runs and reserve triangles into consistent, comparable formats. Cedant and TPA reports that use different column layouts, development period definitions, and currency conventions are normalized into a unified structure. This eliminates the manual data-wrangling step that typically consumes the first weeks of a diligence engagement.

Cross-document reconciliation

The platform’s cross-document reasoning engine reconciles paid, incurred, and reserve movements across all sources in the portfolio. When a bordereaux reports a different incurred figure than the corresponding loss run, or when reserve movements in an actuarial report do not align with TPA data, Parsewise flags the discrepancy with full source attribution (document, page, and paragraph references). This is not retrieval-based matching; the engine processes every page exhaustively, so anomalies in the long tail of a portfolio are not missed.

Structured outputs for pricing and reporting

Extraction results are delivered as structured, schema-based data, not freeform summaries. Outputs map directly to the formats that pricing actuaries, investment committees, and regulators require. Every extracted value links back to its source, making outputs auditable and defensible.

Example Inputs and Outputs

Inputs

  • Cedant loss triangles (PDF and Excel, multiple formats per cedant)
  • TPA loss runs and bordereaux
  • Actuarial reserve reports and exhibits
  • Policy wordings and endorsements
  • Claims correspondence and case notes
  • Regulatory filings and scheme documentation

Outputs

Output Description
Aligned reconciliation tables Standardized loss triangles and reserve summaries across all cedants and TPAs, in a single comparable format
Variance and reserve shift detection Flagged discrepancies between reported reserves, paid amounts, and incurred figures across sources, with exact source citations
Portfolio-level exposure tables Aggregated severity, frequency, and exposure data segmented by line of business, geography, and development period
Red flag reports Structured summaries of data gaps, anomalies, and inconsistencies for investment committee or regulatory review
Benchmark-ready datasets Cleaned, structured data suitable for actuarial modeling, pricing analysis, and regulatory reporting templates

Customer Evidence: Compre Group

Compre Group, a specialist in legacy insurance and reinsurance portfolio acquisitions, uses Parsewise to perform diligence on acquired portfolios and reconcile claims data across multiple sources.

Legacy portfolios typically arrive as large, fragmented document packages spanning loss runs, bordereaux, actuarial reports, and policy documents in varying formats. Parsewise ingests these heterogeneous document sets and automatically standardizes loss runs and reserve triangles into consistent, comparable formats. The platform reconciles paid, incurred, and reserve movements across cedants and TPAs, flagging anomalies, reserve shifts, and data gaps. Compre Group uses the structured outputs for pricing decisions, reserve adequacy assessments, and regulatory reporting on acquired portfolios.

The platform’s ability to process 25,000+ pages per run and maintain context across the entire corpus means that diligence teams can analyze a full portfolio in a single pass rather than sampling subsets manually.

Why Not Sample Manually?

Manual sampling is the default for large portfolio diligence. Analysts select a subset of claims files or cedant reports, review them in detail, and extrapolate findings to the full portfolio. This approach has a fundamental limitation: it cannot guarantee that the sample captures the severity patterns, data gaps, or reserve discrepancies that matter most.

Parsewise processes every page in the diligence package. The Parsewise Data Engine coordinates extraction across the full corpus autonomously, running for over 5 hours per run and handling 20,000+ requests per minute. There are no retrieval gaps, no sampling bias, and no documents dropped because they did not match a query. For a detailed explanation of why retrieval-based approaches fail for this category of work, see Why RAG Fails for Risk-Grade Decisions.

Ready to see Parsewise in action? Request a demo or contact sales to discuss your use case.

Sources