Bordereaux Intelligence
Bordereaux Validation Software — Row-Level Rules Engine for Reinsurance.
CedeOS catches the five systematic error categories that manual review misses — before submission, not after the debit note.
| Policy | Cedant | Class | Cession | Amount | Category | Impact | Status | |
|---|---|---|---|---|---|---|---|---|
| CA-LIF-00441 | Insurance A | Life Credit | 30% | USD 563 | CESSION RATE | USD 217 | FAIL | |
| CA-GEN-01882 | Insurance A | Motor Commercial | 40% | USD 1,692 | EXCLUSION | USD 1,692 | FAIL | |
| CA-LIF-00229 | Insurance A | Life Credit | 34.6% | USD 892 | REINSTATEMENT | USD 31 | REVIEW | |
| CA-GEN-02014 | Insurance B | Property | 35% | USD 2,140 | — | PASS | ||
| CA-GEN-01773 | Insurance B | Marine Cargo | 30% | USD 780 | — | PASS | ||
| CA-GEN-03301 | Insurance C | Property | 35% | USD 3,400 | FX CONVERSION | USD 48 | REVIEW |
The Condition
The bordereaux you submit reflects data from October 10. The policy cancelled on October 14.
Your quarterly submission contains 3,247 rows across 14 cedants. The totals reconcile. The column headers match. The cession rates appear correct when summed by class of business. Your senior actuary signs off.
Week 8: the reinsurer's debit note arrives KES 2.3M short of expectation. The dispute begins. 14 weeks later, you recover at 40–60 cents on the dollar. The root cause: 23 rows across three cedants carried systematic errors invisible at aggregate level.
A vehicle coded as property and ceded at 40% instead of 30%. A cancelled policy still on the bordereaux from a PDF submission by your largest MGA. A rounding pattern on currency conversion that only one analyst noticed — by chance, not by process.
Manual validation cannot cover every row. It checks totals. The errors live in the rows.
Without CedeOS
3,247 rows received
Totals reconcile — senior actuary signs off
Spot-check on aggregate totals
23 row-level errors pass undetected
Submission sent to reinsurer
Debit note arrives USD 1.9M short of expectation
14-week dispute cycle
Recovery at 40–60 cents on the dollar
14 weeks · 40–60¢ recovery
With CedeOS
3,247 rows ingested
Any format — auto-normalised, currency-converted
Every row validated against treaty
23 deviations flagged across 5 error categories
Exceptions reviewed — 1 click
Actuary approves or corrects. Clause cited on each.
Submission sent — dispute-proof
Zero shortfall. Zero dispute. Full audit trail.
Same day · 100% recovery
The Cost
The errors are invisible in aggregate. They live in the rows.
Premium leakage at the row level
Industry average: 2.5% of GWP leaks through systematic row-level errors that aggregate to correct totals. At $50M GWP, that is $1.25M per year — invisible without row-level validation.
2.5% GWP leakage — industry benchmark across African and Asian programmes
14-week post-submission disputes
When errors surface at the reinsurer, the dispute cycle averages 14 weeks. Recovery comes at 40–60 cents on the dollar. The analyst who could have caught it spent their time preparing the submission manually.
14-week dispute cycle — industry baseline across African reinsurance programmes
IFRS 17 audit exposure
The external auditor traces the CSM back to source bordereaux. The file is 'Q2_recon_v4_FINAL_actual.xlsx.' There is no validation trail, no clause citation, no evidence of systematic checking. The finding is material.
IFRS 17 requires full traceability — effective since January 2023
The Standard
What a validated reinsurance submission looks like.
Quarterly close
vs. 6-week manual baseline
Post-submission disputes
Target with pre-submission validation active
Error categories caught per submission
All systematic — none visible without row-level check
Row coverage
Every row validated, not spot-checked totals
Projected Outcome
What row-level validation eliminates.
Pre-submission validation across five error categories eliminates the conditions that create post-submission disputes. When every row is validated against its applicable treaty clause before submission, the reinsurer receives accurate figures. The dispute does not start.
Cession rate misclassification caught
Per reinstatement miscalculation prevented
Currency conversion errors corrected
Per aggregate breach detected early
Actuary time saved per dispute avoided
Treaty structure analysis, CedeOS 2024. Pilot validation in progress. Assessment available on request.
The Mechanism
Five error categories. Each invisible at aggregate level. Each caught at the row.
The CedeOS validation taxonomy was developed from analysis of East African and Asian bordereaux structures and published treaty terms across target markets. Each category has defined detection rules, treaty clause mapping, and financial impact quantification.
Cession Rate Misclassification
A vehicle coded as property and ceded at 40% instead of 30% looks correct in aggregate when mixed with legitimate property cessions. It is wrong on every row. Multiplied across 200 policies, it generates KES 10M per quarter in under-ceded premium.
KES 10M/quarter — CedeOS internal analysis, 2024
Reinstatement Premium Miscalculation
Treaty specifies reinstatement on original annual premium. Endorsement changes the current premium mid-term. System calculates on post-endorsement figure. The clause on page 47 is never checked against each policy's endorsement history.
KES 110K per XL trigger event — treaty structure analysis, CedeOS 2024
Wrong-Date Currency Conversion
Treaty requires conversion at transaction date. System uses month-end rate. The difference compounds across hundreds of cross-border policies. USD 34K per quarter — invisible until the reinsurer's own reconciliation catches it.
USD 34K/quarter — currency analysis across cross-border programmes
Undetected Aggregate Retention Breach
XL treaty retention is KES 50M. Accumulation tracked in a spreadsheet. Two claims in weeks 10-11 push past threshold. Nobody knows until annual audit. Recovery at 40-60 cents — 8 months later.
KES 4M+ per aggregate breach — East African XL programme modelling
Cancelled Policies on Bordereaux
Policy cancelled October 14. Bordereaux data reflects October 10. The MGA's PDF submission carries the policy for the full quarter. Three weeks of dispute per occurrence when the reinsurer's system flags the overlap.
3 weeks per dispute — industry baseline for cancelled policy disputes
35% — Article 3 Medical
30% submitted
Motor Commercial >3.5t excluded — Art.9(b)
Class 06 submitted as Property
Original annual premium at inception — Art.14.2
Post-endorsement premium used
USD 385K net retention — Art.22.1
USD 416K accumulated — not flagged
Active cover period only — Art.6
Policy cancelled Oct 14, Q3 submission Oct 10
Property 35% — Article 3
35% — matches exactly
The Pipeline
Ingest. Validate. Submit. One pass.
Ingest
Any format accepted. Email, SFTP, CSV, PDF, API. 40+ column naming conventions normalised. Currency converted at transaction-date rates. Confidence-scored per field.
Validate
Every row checked against applicable treaty clauses. Five error categories. Cession rates, reinstatements, currency, aggregate accumulation, cancellations. Single pass. Clause cited on every deviation.
Submit
Validated rows submitted automatically. Deviations held for actuary review. One click to approve exceptions. Full audit trail. Submission-ready output for every framework.
40+ column naming conventions · CBK/RBA FX rates · Confidence-scored per field
1,847 rows · 99.3% confidence
Every row × every applicable treaty clause · Clause citation on every deviation
3 FAIL · 2 REVIEW · 1,842 PASS
Validated rows auto-submit · Exceptions to actuary queue · Full audit trail
Zero disputes · 100% traceable
Regulatory Compliance
Validated output for every framework your operation requires.
IRA Kenya
Bordereaux validation against IRA treaty reporting requirements. Cession rate accuracy per class. Quarterly submission format compliance.
NAICOM Nigeria
40% minimum cession rate validated per row. Error quantification in NGN. Regulatory reporting format output.
SAMA Saudi Arabia
Takaful bordereaux structure validated. Wakala fee accuracy per participant. Conventional and Islamic programme support.
IFRS 17
Every validation decision logged with clause citation. CSM impact of errors quantified. Full traceability from source policy to financial statement.
The Path
From first assessment to validated submission in four weeks.
Day 0
Technical Brief
Architecture, five error categories, validation methodology, data residency, IFRS 17 audit trail approach.
Week 1
Pilot Assessment
One quarter of bordereaux data. We return a validation report showing errors by category, cedant, and financial impact.
Week 2
Live Demo
Your data, your treaties. See exactly what the five error categories surface in your own submission.
Week 4
First Validated Submission
Pilot go-live. First bordereaux validated row by row. Every deviation clause-cited. Submission dispute-proof.
Get started