Multi-Channel Ingestion
Bordereaux Ingestion Software — Any Format, One Pipeline.
40+ column naming conventions. Five ingestion channels. Takaful separation at ingestion. Your cedants send however they want — CedeOS normalises it automatically.
Why 72% — actuary confirmation needed
"CESSION_%" matches cession rate with high probability. However, 3 other fields in this file contain "%" symbols. CedeOS selected cession_rate as most likely — confirm or remap.
The Condition
Your largest cedant sends a PDF. Your smallest sends a CSV with columns in Arabic.
You have 14 cedants. Each sends bordereaux in their own format. Your analyst spends 35 hours per week converting these into something your system can read. A new cedant takes 2-4 weeks to onboard because the format mapping is manual.
The Takaful MGA sends a mixed bordereaux with Wakala and Mudaraba risks unseparated. Your broker in Nairobi uses a format nobody has seen before. The Gulf cedant has column headers in both Arabic and English. All of this is your analyst's job — at $70K per year.
This is an infrastructure problem with a measurable cost. Not a people problem.
All of these mean "gross premium" — 40+ naming conventions from real cedant submissions:
CedeOS internal cataloguing, 2023-2024. African, Middle Eastern, and Asian cedant formats.
The Cost
Format conversion is an infrastructure cost — not a people cost.
Analyst cost recoverable per year
1 FTE fully-loaded. MGA operations baseline.
Analyst hours/week on format conversion
MGA operations baseline; industry interviews.
Column naming conventions per field
CedeOS internal cataloguing, 2023-2024.
Per new cedant onboarding
Manual format mapping exercise.
Without CedeOS — Weekly analyst schedule
35 hrs/week · $70K/year · zero analysis
With CedeOS — Weekly analyst schedule
< 2 hrs/week on formats · $70K/yr recaptured
The Standard
What ingestion looks like when it works.
Cedant sends whatever format they have. No reformatting requested.
Column headers mapped automatically. 40+ conventions per field normalised.
Takaful and conventional risks separated at ingestion — not manually later.
Validation engine receives clean canonical data. Every time. First time.
Projected Impact
Format conversion cost is quantifiable. Elimination is measurable.
Current state (manual)
Analyst FTE on format conversion
Manual normalisation and reformatting
Per new cedant format mapping
Target with CedeOS
Analyst time on format-related exceptions
Manual format conversion required
Takaful/conventional separation
MGA operations baseline. Industry interviews and internal analysis. Pilot validation in progress.
Five Channels
Bordereaux arrive however your cedants send them.
Excel / CSV
Any column order, any naming convention, any language. Mapped automatically with confidence scoring per field.
REST API
JSON/CSV payloads via webhook or scheduled pull. System-to-system for modern policy administration systems.
Table extraction from scanned and digital PDFs. Bordereaux that arrive as attachments from MGAs who only produce paper.
Forwarded directly to a CedeOS inbox. Attachments extracted, parsed, and normalised. The cedant changes nothing.
SFTP
Scheduled file drops. Legacy systems that produce flat files overnight. Picked up, parsed, normalised before your team arrives.
All channels → one pipeline
CedeOS Normalisation Engine
Canonical output · Every time
Takaful Native
Wakala and Mudaraba separated at ingestion. Not manually. Not later.
Gulf cedants submit mixed bordereaux containing Takaful (Wakala and Mudaraba) and conventional risks in the same file. Most platforms require manual separation before processing. CedeOS separates at ingestion — automatically.
Wakala/Mudaraba identification per row based on product classification
Mixed bordereaux → distinct validated submissions per pool
Regulatory treatment applied per pool (SAMA Takaful cession rates)
No configuration required — architecture-level separation
Takaful Separation — Automatic at Ingestion
Mixed bordereaux received
insurance-a_Q3.xlsx · 412 rows
CedeOS
Separates at ingestion
Three validated outputs
Wakala Pool
168 rows · 30% cession
Mudaraba Pool
96 rows · 28% cession
Conventional Pool
148 rows · 35% cession
SAMA Takaful cession rates applied per pool
Normalisation
Receive. Normalise. Validate. One pass.
Receive
File arrives via any channel. Format detected. Column headers identified. Language detected (English, Arabic, French, Portuguese).
Normalise
Each column mapped to canonical field. Confidence score per mapping. Currency identified and converted at transaction-date rate. Takaful separation applied.
Validate
Normalised data enters validation pipeline. Every row checked against applicable treaty clauses. Deviations flagged with confidence < threshold held for review.
Field Mapping — Confidence Scores
1 field needs confirmation| Source Column | Canonical Field | Confidence | Action |
|---|---|---|---|
| PREM_AMT | gross_premium | 99% | Auto-mapped |
| POLICY_REF | policy_number | 98% | Auto-mapped |
| INCEPTION_DT | inception_date | 97% | Auto-mapped |
| CESSION_% | cession_rate | 72% | |
| RISK_CLASSIFCTN | risk_class | 91% | Auto-mapped |
CESSION_% matched cession_rate at 72% — 3 other fields contain "%" in this file. CedeOS selected most likely match. Actuary confirmation locks this mapping for all future submissions from this cedant.
The Path
From sample files to production ingestion in four weeks.
Day 0
Technical Brief
Architecture, supported formats, normalisation methodology, Takaful handling.
Week 1
Format Mapping Report
Submit sample files from your cedants. We return a mapping report: columns identified, confidence scores, exceptions flagged.
Week 2
Live Demo
Your cedant files, ingested live. See confidence scores, Takaful separation, and normalised output in real time.
Week 4
Production Ingestion
All cedant formats onboarded. Automated pipeline active. Analyst recaptured for high-value work.
Get started