Bordereaux, Treaty Validation, and Reinsurance Compliance — Answered.

A bordereaux (plural: bordereaux) is a detailed report listing individual policies or claims that an insurer or MGA submits to a reinsurer. It contains row-level data — policy numbers, premiums, cession rates, risk classes, and claims — used for reconciliation, premium accounting, and regulatory reporting. In African and Asian markets, bordereaux arrive in 40+ different column naming conventions across Excel, CSV, PDF, and email formats.

Bordereaux validation is the process of checking every row in a bordereaux submission against the applicable treaty clauses before it reaches the reinsurer. This includes verifying cession rates, reinstatement calculations, currency conversions, aggregate retention limits, and policy cancellation status. Without automated row-level validation, errors are invisible at aggregate level and discovered only during post-submission disputes — typically 3 to 18 months later.

A bordereaux rules engine is software that automatically applies treaty-defined rules to every row in a bordereaux submission. It extracts clauses from reinsurance treaty documents (retention limits, reinstatement conditions, exclusions, aggregate caps) and validates each row against the applicable rules. CedeOS extracts these rules automatically from full treaty documents and applies them at ingestion — catching errors before submission rather than after the debit note.

Bordereaux data standardization is the process of normalising different column naming conventions, data formats, currencies, and structures into a canonical schema that can be validated and reported against. Cedants across Africa and Asia submit data in 40+ different naming conventions for the same field (e.g., "PREM_AMT", "Premium", "GWP", "Gross Premium LCY" all mean gross premium). CedeOS normalises these automatically with confidence-scored AI mapping.

Premium leakage is the systematic loss of reinsurance premium revenue caused by row-level errors in bordereaux data that are invisible when checking aggregate totals. The industry average is 2.5% of Gross Written Premium (GWP) per year. For an insurer writing $50M GWP, that is $1.25M annually — not from fraud, but from process failure. The five main causes are: cession rate misclassification, reinstatement premium miscalculation, wrong-date currency conversion, undetected aggregate retention breaches, and cancelled policies remaining on submissions.

CedeOS validates every bordereaux row against the applicable treaty clauses at ingestion — before any figure enters the accounts or reaches the reinsurer. Each discrepancy is quantified in local currency and USD, cited back to the exact treaty clause, and categorised by error type. This is prospective detection: errors are caught on the same day the data is submitted, not 3-18 months later when the reinsurer raises a dispute.

Reinsurance recovery for leaked premiums involves identifying historical bordereaux errors, quantifying their financial impact per cedant and treaty clause, and generating evidence documentation for recovery discussions. CedeOS supports both prospective detection (catching errors before submission for full recovery) and retrospective recovery (re-validating historical portfolios to identify past leakage). Retrospective analysis typically surfaces 2.5% of GWP in recoverable leakage across five error categories, with recovery documentation generated automatically — including the incorrect value, correct value, governing clause, and financial impact per cedant.

Treaty validation is the process of verifying that every row in a bordereaux submission complies with the rules defined in the applicable reinsurance treaty. This includes checking cession rates per risk class, applying reinstatement premium calculations per the treaty schedule, enforcing named peril and geographic exclusions, and monitoring aggregate retention limits. Without automated treaty validation, these rules are applied inconsistently by different analysts across quarters — creating IFRS 17 CSM variance and post-submission disputes.

CedeOS parses full treaty documents (typically 40-80 pages) and extracts structured rules: retention limits, reinstatement conditions, exclusions, cession rates, aggregate caps, and currency requirements. These rules are indexed and applied row-by-row at ingestion. Every deviation is flagged with the specific clause reference, the incorrect value, the correct value according to the treaty, and the financial impact in local currency.

Traditional aggregate monitoring relies on quarterly spreadsheet reconciliation — meaning breaches are discovered months after they occur. CedeOS monitors aggregate accumulation continuously as each claims batch processes. An alert is generated at 80% of the retention threshold with the current figure, clause reference, and projected breach date. The threshold is never silently crossed.

Reinstatement premium errors occur when the premium charged to reinstate coverage after a loss event is calculated on the wrong basis. The treaty typically specifies reinstatement on the original annual premium, but mid-term endorsements change the current premium. Systems that calculate on the post-endorsement figure rather than the original contract premium create systematic errors — typically KES 110K per XL trigger event. These compound every quarter and are invisible without row-level validation against the specific treaty clause governing reinstatement.

IFRS 17 requires full traceability from source policy to financial statement — specifically for the Contractual Service Margin (CSM). CedeOS generates a complete audit trail as a natural output of the validation process: every transformation is logged with input value, output value, governing rule, treaty clause reference, and confidence score. When the external auditor asks to trace a CSM figure back to source, the answer is one click — not three weeks of shared-drive archaeology.

CedeOS automates compliance across 8 frameworks: IRA Kenya (treaty reporting, cession rate validation), NAICOM Nigeria (40% minimum cession, local currency reporting), FSCA South Africa (prudential compliance, POPIA data residency), SAMA Saudi Arabia (Takaful and conventional, in-kingdom processing), SECP Pakistan (treaty filing, local currency), DIFC UAE (cross-border treaty reporting), IFRS 17 (CSM/Loss Component, full audit trail), and POPIA (South African data protection).

CedeOS processes data regionally — not as a configuration option, but as an architectural guarantee. African data is processed in af-south-1 (Cape Town). Gulf and Asian data is processed in me-central-1 (Qatar/UAE region). No cross-border data egress occurs by default. This satisfies POPIA (South Africa), SAMA (Saudi Arabia), and IRA (Kenya) data residency requirements without additional configuration.

Under IFRS 17 (effective January 2023), reinsurers and insurers must maintain a complete transformation chain from source policy to financial statement — specifically for CSM and Loss Component calculations. This means every bordereaux row must be traceable through normalisation, validation, treaty application, and reporting. Additionally, IRA Kenya, NAICOM Nigeria, and SAMA Saudi Arabia each require evidence of correct treaty application in their respective reporting formats. CedeOS generates this audit trail as an automatic side-effect of processing — using SHA-256 hash chains for tamper-evident evidence with 7-year retention.

Excel cannot validate individual bordereaux rows against treaty clauses — it can only check totals. This means systematic errors (wrong cession rates, reinstatement miscalculations, cancelled policies) are invisible because they aggregate to correct-looking totals. Additionally, Excel has no audit trail, no version control that satisfies IFRS 17, no real-time aggregate monitoring, and no mechanism to apply 80-page treaty rules programmatically. The industry average: 2.5% of GWP leaks annually through errors that pass every manual Excel spot-check.

Bordereaux reconciliation is the process of matching the data in a bordereaux submission (individual policy rows, premiums, claims) against treaty terms, premium accounts, and financial statements. Traditionally done in Excel over 6 weeks per quarter, it involves: verifying cession rates per risk class, checking currency conversions, confirming reinstatement calculations, validating aggregate retention status, and producing submission-ready output for the reinsurer. CedeOS automates this entire process — targeting under 3 hours for a complete quarterly close with full audit trail.

Manual reconciliation: 6 weeks per quarter, spot-checked totals only, no audit trail, 14-week dispute cycle when errors surface, 40-60 cent recovery on disputed amounts. CedeOS: under 3 hours, every row validated against every applicable clause, complete IFRS 17 audit trail, zero post-submission disputes (errors caught before submission), 100% recovery (errors never leave your system). The infrastructure cost of CedeOS is recoverable from a single quarter of prevented leakage.

Reinsurers evaluate MGA bordereaux submissions against several key benchmarks: cession rate accuracy (must match treaty schedule per risk class), currency consistency (conversion at transaction-date central bank rates, not month-end), timeliness (submission within treaty-defined reporting windows), policy status accuracy (no cancelled or lapsed policies on active submissions), and format compliance (column mapping to reinsurer-specified schema). MGAs that consistently fail these benchmarks face longer settlement cycles, disputed debit notes, and damaged carrier relationships. CedeOS validates all of these automatically before carrier submission — reducing query cycles and accelerating settlements.

Traditional reinsurance back-offices rely on manual Excel reconciliation, spot-checked totals, and quarterly human review. AI transforms this by enabling: reasoning-based treaty validation (reading full 80-page treaty documents and applying clauses row-by-row), automated bordereaux normalisation (mapping 40+ column naming conventions across languages without analyst intervention), continuous aggregate monitoring (real-time threshold alerts instead of quarterly spreadsheet checks), and automated audit trail generation (every decision logged with clause citation and confidence score). CedeOS uses verticalized AI specifically trained on reinsurance treaty structures and African/Asian regulatory requirements — not general-purpose document scanning.

CedeOS targets first validated output within the first engagement week. The typical path: Day 0 (technical brief), Week 1 (pilot assessment with one quarter of your bordereaux data), Week 2 (live demo with your actual data), Week 4 (first validated quarterly close in production). The team has direct African and Asian market experience — no London-first learning curve.

Submit one quarter of bordereaux data from your top cedants. CedeOS returns a validation report showing: errors found per cedant, financial impact per error category, treaty clause violations, projected leakage exposure, and compliance gaps. There is no commitment. The assessment itself demonstrates the value before any contract discussion.

CedeOS integrates alongside existing policy administration systems via REST API, SFTP, or email ingestion. It does not require replacing your core systems. Bordereaux arrives however your cedants currently send it — the format conversion and validation layer sits between your existing process and the reinsurer submission.