A finance director at a mid-size reinsurer pulls up two numbers during an annual budget review. The first is the SICS platform cost — the license, the hosting, the DXC support contract. It is substantial but predictable. The second is the fully loaded cost of the team that feeds data into SICS — the operators who read bordereaux PDFs, interpret Excel tabs, and manually rekey entry codes every day.
That second number is five to eight times the first.
The platform is not the expense. The translation is.
DXC's Strategic Information and Cession System runs the reinsurance back office at over 100 reinsurers in 38 countries. It is a mature, powerful system of record for treaty management, cession accounting, and financial settlements. What SICS does not do — and has never done — is read the unstructured documents your cedants actually send you.
Your cedants send bordereaux as PDFs. Excel files with inconsistent column layouts. Emails with data pasted inline. Scanned cover notes. Narrative adjustment letters that span three prior quarters. SICS expects structured entry-code data. Your operators bridge the gap. And that bridge carries a cost far higher than most reinsurers calculate — because nobody accounts for the translation tax.
A reinsurer with 50 cedants processes a minimum of 300 entry decisions a month — written premium, commission, brokerage, paid loss, case reserves, premium adjustments — across every reporting period. A reinsurer with 200 cedants is closer to 1,200 entry decisions a month, each one requiring a human to read an unstructured document, interpret it in the context of that specific cedant's quirks, and rekey it into SICS.
McKinsey and Accenture estimate that 30-40% of underwriter and operations time in insurance goes to administrative tasks. In reinsurance cession-accounting teams, that percentage is almost certainly higher — because the unstructured-to-structured translation is the core of the job, not a side task.
Here is a rough cost envelope for a mid-size reinsurer's translator desk:
The translation tax runs five to eight times the platform cost. Unlike the platform license, it scales linearly with the number of cedants. Every new program, every new treaty, every new cedant dialect adds work. The team grows with the book — or the team stays the same size, and accuracy erodes.
DXC Technology reported $12.9 billion in revenue for fiscal 2025 — down 5.8% year-over-year. Its headcount has declined from 150,000 in 2018 to 120,000 in 2025. This is a company optimizing its existing portfolio, not building new capabilities at the edge of its platform.
DXC's answer to the data-exchange problem is the ACORD ADEPT integration — a partnership that enables structured digital exchange between brokers, cedants, and reinsurers who have both adopted the standard. Digital reinsurance transactions on ADEPT grew 73% from 2024 to 2025, which sounds impressive until you ask what share of your cedant base has actually adopted it.
The answer, for most reinsurers, is roughly 20%. The other 80% — the long tail of small and mid-size cedants in every market — still send PDFs, Excel files, and emails. They are not on ADEPT's adoption roadmap for next year. The investment required for each cedant to adopt a structured data-exchange standard is non-trivial relative to their interest in accommodating your specific platform preferences.
ADEPT is real, and it is growing. It solves the structured 20%. But DXC is not building a solution for the unstructured 80%. The gap between what SICS can accept natively and what your cedants actually send is widening, not closing.
Every treaty relationship you add brings another bordereaux dialect — different column layouts, different reporting conventions, different definitions of what "commission" includes. The operator who just mastered Cedant A's format starts from scratch on Cedant B. A reinsurer growing its book by 10-15% a year is adding complexity faster than the team can absorb it.
The experienced operator who knows that Cedant X always reports premium net of brokerage in column F — and that Cedant Y's profit commission true-ups arrive in narrative form two months late every Q3 — carries that knowledge in their head. When they leave, their successor faces a 90-180 day ramp before they stop making mistakes their predecessor had already learned from. With 20-40% annual turnover in insurance back offices, that dictionary leaks constantly.
Cross-jurisdiction tax variance — UK insurance premium tax, EU stamp duties, state-level levies — means the same premium number carries different net values depending on where the risk sits. A cedant reporting a blended number and an operator posting it without the jurisdictional adjustment creates an error that surfaces not this month but at year-end audit.
Each of these forces adds work to the translator desk. None of them are addressed by SICS natively, by ADEPT, or by hiring more operators.
The most expensive moment in a cession-accounting team is not a mis-posted entry. It is the departure of the senior operator.
The senior operator has built a private dictionary over years. They know every cedant's format quirks, exception patterns, late-arriving adjustments, and narrative-form true-ups. That dictionary is the reinsurer's institutional knowledge of how each cedant reports — and it does not live in any system your reinsurer owns. It lives in the operator's head.
When that operator retires, accepts a competitor's offer, or goes on extended leave, the dictionary walks out with them. The replacement — competent, trained on SICS, motivated — starts from zero on every cedant relationship the departing operator had memorized. Errors in the first six months are not a failure of talent. They are the cost of lost memory.
The reinsurance-administration software market is projected to grow from $2.7 billion in 2024 to $7.1 billion by 2033 at a compound annual rate of 10.8%. That growth reflects automation pressure and contract complexity. But the platform investment captures only half the cost picture. The translation tax — the human cost of bridging unstructured documents to structured platforms — is the other half. And it is the half that scales badly.
Brisc AI's Reconciliation Analyst is purpose-built to be the translator that stays. It reads the unstructured documents your cedants send — PDFs, Excel, email, scanned notes — and classifies every field into your specific SICS entry codes using a per-customer dictionary that accumulates institutional knowledge with every cedant interaction.
The first time Brisc encounters a new cedant's bordereaux, your senior operator reviews the classifications and corrects what needs correcting. The dictionary learns. The second time, Brisc handles that cedant correctly without review. By the 50th interaction, the dictionary has absorbed every format variation, every reporting quirk, every late-arriving adjustment pattern that cedant has ever sent.
When your senior operator leaves, that dictionary remains. Their successor starts on day one with the accumulated knowledge of every cedant your reinsurer has ever processed.
Brisc AI maintains 97%+ accuracy on bordereaux reconciliation — the same on the first document of the day as on the thousandth. Brisc customers report a 59% labour cost reduction across operations. One multi-line MGA using Brisc reports an 80% reduction in manual labour on submissions intake. Deployments are live in 2-6 weeks depending on cedant complexity.
Brisc is not a SICS replacement. SICS is the system of record; it stays. Brisc is the translation layer SICS was always missing — the bridge between the documents your cedants send and the structured entries your platform expects. For the six core entry categories that consume most of a cession-accounting desk's day, see Six SICS Entries Your Team Translates by Hand Every Day.
Does Brisc replace SICS?
No. Brisc AI is the translation layer between unstructured cedant documents and SICS. SICS remains the system of record. Brisc reads bordereaux, classifies entry codes, and outputs in SICS upload format. Your team imports the output using the same workflow they use today.
What about ACORD ADEPT — doesn't that solve this problem?
ADEPT solves the structured data-exchange problem for cedants who have adopted the standard — roughly 20% of most reinsurers' cedant base today. Brisc handles the other 80% who still send PDFs, Excel, and email. The two are complementary: ADEPT for the digitized minority, Brisc for the unstructured majority.
How long does deployment take?
Brisc deployments are typically live in 2-6 weeks. The variable is the number of cedant dictionaries to configure at launch. Each subsequent cedant onboarded is faster than the last as the dictionary accumulates shared patterns across your book.
What happens when a cedant changes their bordereaux format?
Format drift — when a cedant silently changes column definitions or reporting conventions — is one of the highest-cost problems on a cession-accounting desk. Brisc detects format changes automatically, flags them for review, and updates the dictionary once the operator confirms the new pattern. The next occurrence processes without intervention.
Is Brisc only for SICS reinsurers?
No. The architecture works with any reinsurance administration platform — Sapiens ReinsuranceMaster, Verisk Sequel, in-house systems. The engine is the same; the entry-code dictionary and output template change per platform. SICS is the largest deployed platform, which is why this post focuses on it.
Can we start with a subset of cedants?
Yes, and most reinsurers do. A typical deployment begins with the highest-volume cedants — the ones consuming the most operator time — and expands from there. Each cedant added to the dictionary makes the next one faster.
How do we know the classifications are correct?
Every classification ships with a confidence score and a complete source-document reference. Low-confidence entries route to your operator for human review. High-confidence entries process automatically. The full audit trail — source document, extracted fields, classification logic, resulting SICS entry — is captured continuously, not reconstructed after the fact.
What does "97% accuracy" mean in practice?
It means 97 out of every 100 entry-code classifications Brisc produces match the classification your senior operator would have made, verified across production deployments. The remaining 3% are routed for human review. That accuracy holds consistently regardless of volume or time of day — the structural difference from a human desk where accuracy degrades as fatigue sets in.
If you are running SICS, your team spent today translating. The platform cost is fixed and predictable. The translation cost scales with every new cedant, every new treaty, every staff departure that erases a piece of the dictionary.
The dictionary should live in the system, not in someone's head.
See how Brisc's Cession Analyst handles your cedant bordereaux