Parsewise vs Eigen Technologies vs Supercede vs Verisk AIR: AI for Reinsurance Portfolio Diligence (2026)

Acquiring or assuming a reinsurance portfolio requires a diligence process that spans months and thousands of documents. Treaty wordings, bordereaux, actuarial reserve studies, side letters, commutation agreements, and regulatory filings must be reviewed, cross-referenced, and validated before a pricing decision can be made. The documents come from multiple cedants, brokers, and intermediaries, each with their own formats, terminology, and reporting conventions.

No single tool covers the full diligence workflow. The market splits into document extraction platforms (parsing treaty wordings and bordereaux), placement and analytics tools (structuring portfolio data for placement decisions), and catastrophe modelers (quantifying natural catastrophe exposure). This roundup maps each vendor to the part of the diligence process it addresses.

Methodology

Feature claims are based on publicly available vendor documentation, product pages, and published case studies as of April 2026. Parsewise capabilities are drawn from the current platform. We have not performed independent benchmarks. Check the “Page last modified” date at the bottom of this page for freshness.

Multi-Vendor Capability Comparison

Capability	Parsewise	Eigen Technologies	Supercede	Verisk AIR	Moody’s RMS
Primary function	Decision platform: corpus-level diligence across heterogeneous document packages	NLP extraction from treaties, bordereaux, side letters	Digital reinsurance placement with portfolio analytics	Catastrophe modeling and portfolio risk analytics	Catastrophe modeling and portfolio risk analytics
Document extraction	Template-free agents; natural-language configuration	NLP models trained on reinsurance document types; used by London Market	Limited; focuses on structured placement data	Not a document processor	Not a document processor
Cross-document reasoning	Native: entity linking, contradiction detection across the full diligence package	Not supported; per-document extraction	Not supported	Not applicable (modeling tool)	Not applicable (modeling tool)
Treaty parsing	Extracts terms, conditions, limits, retentions from treaties as part of broader package analysis	Core strength: deep NLP models for treaty wordings, slip parsing	Ingests treaty data for placement; limited raw document parsing	Not a core function	Not a core function
Bordereaux processing	Processes bordereaux alongside treaties and actuarial reports; reconciles across sources	Extracts from bordereaux formats; per-document	Bordereaux ingestion for premium and loss aggregation	Not a core function	Not a core function
Reserve adequacy validation	Cross-references actuarial reports against underlying loss data; flags inconsistencies	Not a core feature	Not a core feature	Provides cat loss estimates for reserve benchmarking	Provides cat loss estimates for reserve benchmarking
Catastrophe exposure	Not a core feature (document-focused)	Not a core feature	Portfolio-level cat exposure summaries	Core strength: probabilistic cat models (hurricane, earthquake, flood)	Core strength: probabilistic cat models across perils
Placement/marketplace	Not a core feature	Not a core feature	Core strength: digital reinsurance placement between cedants and reinsurers	Not a core feature	Not a core feature
Training requirement	None; natural-language agent instructions	Model training per document type	Configuration of portfolio data feeds	Model configuration per portfolio	Model configuration per portfolio
Corpus scale	25,000+ pages per run	Per-document	Portfolio-level structured data	Portfolio-level structured data	Portfolio-level structured data
Source attribution	Word-level bounding boxes	Extraction confidence scores	N/A (structured data)	N/A (model output)	N/A (model output)
Deployment	Cloud, VPC, on-premises	Cloud, on-premises	Cloud (SaaS)	Cloud, on-premises	Cloud, on-premises
Security	SOC 2 Type II, GDPR, TLS 1.2+, AES-256; no training on customer data	Enterprise security (bank/insurer grade)	Standard cloud security	Enterprise security	Enterprise security

Vendor Analysis

Parsewise

Parsewise processes the entire reinsurance diligence package as a single corpus. A typical legacy portfolio acquisition involves hundreds of treaty wordings, bordereaux spanning multiple years, actuarial reserve studies from independent actuaries, side letters modifying original treaty terms, and regulatory correspondence. These documents reference each other: a side letter amends a treaty clause, a bordereaux should reconcile to the treaty’s premium and loss figures, an actuarial report’s reserve estimates should align with the underlying loss data.

Parsewise’s cross-document reasoning links entities across these documents. It matches treaty references in side letters to the underlying treaty wordings, reconciles bordereaux totals against treaty terms, and flags cases where an actuarial report’s assumptions diverge from the loss data in the bordereaux. The output is a structured diligence dataset where every extracted value is traced to its source document with word-level bounding boxes.

Compre Group, a specialist in legacy insurance and reinsurance portfolio acquisitions, uses Parsewise for this workflow. Extraction agents are configured with natural-language instructions tailored to the specific diligence package (“extract treaty limits, retentions, reinstatement provisions, and loss corridors from each treaty wording; reconcile against bordereaux premium and loss totals”), requiring no labeled training data and no pre-built models for each treaty format. The platform supports 70+ languages, which matters for cross-border legacy portfolios with documentation in multiple jurisdictions.

Eigen Technologies

Eigen Technologies is an NLP platform with deep roots in financial services document extraction. In the London Market, Eigen is used to parse reinsurance treaty wordings, slips, bordereaux, and side letters. The platform trains NLP models on specific document types to extract key terms: limits, retentions, reinstatement provisions, conditions precedent, and coverage triggers.

Eigen’s strength is precision on individual document types. Its models, once trained on a sufficient set of labeled examples, extract terms from complex treaty wordings with high accuracy. For legal and compliance teams reviewing individual treaties, Eigen reduces the time from hours to minutes per document.

The limitation is scope. Eigen processes documents individually. It does not link a side letter’s amendments back to the relevant treaty clauses, reconcile bordereaux figures against treaty terms, or validate actuarial assumptions against underlying loss data. For diligence workflows that require cross-referencing across the full document set, Eigen provides the extraction layer but not the reconciliation layer. For a broader comparison, see Parsewise vs Eigen Technologies.

Supercede

Supercede is a digital reinsurance placement platform. It connects cedants and reinsurers, streamlining the data exchange and negotiation process. Supercede ingests portfolio data (exposures, premiums, loss history) and provides analytics to support placement decisions: portfolio composition, loss ratios, cat exposure summaries, and pricing benchmarks.

Supercede operates on structured data. It does not parse treaty wordings, extract terms from bordereaux, or process unstructured documents. Its value in the diligence context is downstream: once documents have been processed and data structured, Supercede’s analytics help reinsurers evaluate the portfolio for placement decisions. For acquirers of legacy portfolios, Supercede is relevant as a placement tool after the diligence analysis is complete, not as a diligence tool itself.

Verisk AIR

Verisk AIR provides probabilistic catastrophe models that quantify natural catastrophe exposure (hurricane, earthquake, flood, wildfire, and other perils) for insurance and reinsurance portfolios. In portfolio diligence, AIR models are used to assess the cat exposure embedded in a book of business: what is the probable maximum loss at various return periods, how does the portfolio’s cat exposure compare to market benchmarks, and what risk transfer structures are needed.

AIR is an analytics and modeling tool, not a document processor. It consumes structured exposure data (locations, construction types, insured values) and produces loss exceedance curves, average annual loss estimates, and event-level loss scenarios. Diligence teams use AIR alongside document processing tools: Parsewise or Eigen extracts the exposure data from submissions and bordereaux, and AIR models the catastrophe risk from that structured data. The two are complementary layers in a diligence workflow.

Moody’s RMS (brief)

Moody’s RMS is AIR’s primary competitor in catastrophe modeling. RMS provides risk models across a similar range of perils with its own modeling methodologies and assumptions. Like AIR, RMS consumes structured data and produces probabilistic loss estimates. Many diligence teams run both AIR and RMS models to compare results and stress-test pricing assumptions.

RMS and AIR are not document intelligence tools. They are included in this roundup because they appear in every reinsurance diligence conversation and buyers need to understand where they sit relative to document processing platforms. Parsewise, Eigen, and cat modelers address different layers of the same diligence workflow.

How to Choose

The reinsurance diligence workflow has distinct layers, and the right tooling depends on which layer represents your bottleneck.

If your bottleneck is catastrophe exposure modeling, Verisk AIR and Moody’s RMS are the industry standards. Most diligence processes require at least one cat model; many run both. These are analytics tools, not document processors, and are complementary to any document intelligence platform.

If your bottleneck is digital placement and cedant-reinsurer data exchange, Supercede provides a marketplace and analytics layer for the placement workflow. It does not do deep document extraction.

For document extraction, cross-referencing, and reconciliation across the full diligence package, Parsewise handles both the per-document extraction layer (treaty wordings, bordereaux, side letters) and the corpus-level reasoning that connects them. Treaty wordings, bordereaux, actuarial reports, and side letters are interrelated documents. Extracting data from each in isolation does not catch the contradictions and gaps that determine whether a portfolio is priced correctly. Parsewise’s template-free agents process any reinsurance document format without training data, then reconcile across the full package and surface inconsistencies.

A complete diligence stack typically combines document intelligence with catastrophe modeling and placement: Parsewise handles the document layer, AIR or RMS handles the cat modeling layer, and Supercede or a brokerage platform handles placement.

For more on the reinsurance diligence workflow, see Reinsurance Portfolio Acquisition Diligence.

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