Reinsurance submission preparation in days, not weeks — every figure traced, last year’s questions answered
A checked renewal submission in days, not weeks, every figure traced.
See one case, screen by screen ↓Software that builds the submission a cedant sends to the reinsurance market at renewal — program structure, exposure profile, loss history, cat model output and underwriting changes — from the carrier’s own systems. Reinsurance Submission Preparation adds six agents that build each chapter, tie every figure across chapters and to its source, and answer last year’s reinsurer questions in the new pack.
How to choose AI solutions for insurance →
Updated
Before every treaty renewal the ceded team has to put a submission in front of the market: the program structure, the exposure profile, ten years of cat events and large losses, the catastrophe model output and what changed in underwriting. The figures come from policy, claims, modelling and finance systems that do not agree. The TIV in the policy system, the TIV in the modelled set and the TIV in the cover letter are three different numbers.
So an analyst rebuilds the pack in spreadsheets, one chapter at a time, over four to six weeks. Last year’s reinsurer questions sit in a broker email chain and come back at roadshow week, asked again. When a reinsurer finds a figure that does not tie, they load for the uncertainty — and in a falling market, the cedant with the cleanest, best-explained submission gets the best terms.
Estimated split for one property cat renewal submission, by hand and with the solution, in working days.
Six agents build each chapter, tie every figure and answer last year's questions; the head of ceded reinsurance signs.
Harborline Mutual’s 2027 Property Catastrophe Program — $550M excess of $50M in four layers, seven reinsurers on the panel — is due to Callister Re Brokers in 11 days. Here is what happens next, screen by screen, in the working solution.
The pack opens on the program: expiring premium $43.2M, seven reinsurers, broker Callister Re Brokers, due Fri 30 Oct. All 31 source files are in — the policy extract at 30 Sep, the claims listings, the cat model output and own-view note, 14 rate filings and 22 underwriting notes, and last year’s pack with the 18 questions the panel asked. Every chapter reads “Not built · sources in”.
“Nothing is sent to anyone until you release it.”
The run log shows every step as it happens. The exposure profile builder reads the policy extract — 412,860 locations in force at 30 Sep, the Saltmarsh bordereau joined — and builds chapter 2. The loss history builder reads the claims listings, trends them to 2027 on Megan Tsai’s on-level factors, sees Ida develop from $138.9M to $142.6M and finds two large losses that are not on the list.
Chapter 1 fills in question by question. Halvard Re’s Q1 on TIV growth gets its answer — $86.4B against $78.9B, +4.1 points from indexation, +5.4 points from more policies — and a flag to § 2.3 Exposure profile, where a margin note marks it. Fifteen are answered, two partly, one waiting on another chapter.
“Pack assembled · 15 of 18 questions answered up front · 7 gaps with owners”
Every figure is tied across the five chapters, the cover letter and the finance plan at a 0.5% tolerance. Seven gaps come back, one blocking release: the Saltmarsh program is not in the modelled exposure. Subject premium differs between the cover letter ($1.14B) and the finance plan ($1.12B); two large fires are missing from the list; roof age is unknown for 31% of North Carolina homeowners.
“Nothing is fixed by picking a number.”
Chapter 4 draws the occurrence exceedance curve with the four layers on it: 1-in-100 at $512M, program top reached at 1-in-169. But the model input holds $84.1B against an exposure profile of $86.4B. The 6,214 locations written through Saltmarsh Coastal Underwriters since 1 Apr were loaded after the model input was built on 22 Sep — and Corvane Re asked about exactly this last year.
The reconciler never runs or alters the cat model. It shows the gap and asks the cat modelling lead to re-run.
The re-run puts the 1-in-100 at $531M, the 1-in-250 at $711M and the average annual loss at $64.0M; the program now exhausts at 1-in-148, against 1-in-226 in last year’s pack. The bridge shows why the 1-in-100 moved from $468M: exposure growth +$39M, Cape Cod non-renewals −$8M, vendor model update +$6M, own view of risk +$7M, Saltmarsh locations +$19M.
“Re-run back from Hugo · 1-in-100 $531M · Q3 answered · chapter 4 signed”
Renata asks the assistant to add the two missing large losses and tell Corvane Re where — or presses “Add both to the list” on the gap. The warehouse fire in Linden, NJ ($7.8M, claim CL-24-031877) and the marina plaza fire in Bayport, NY ($5.6M, claim CL-25-047209) join the list: 12 losses of $5M or more, and Q10 is answered in § 3.4.
Every number in the pack has a trail. Total insured value, 2027: $86.4B, § 2.1, from the policy administration system, inforce_20260930.csv, 412,860 rows, extracted Tue 13 Oct 06:10, produced by the exposure profile builder. The 1-in-100 traces to Hugo’s own-view EP curves; the 2027 subject premium to the underwriting projection of 9 Oct.
“A figure without a trail cannot be released.”
Before it goes: five chapters and the cover letter built, chapter 4 signed by Dr. Hugo Lindqvist, 1,284 of 1,284 figures with a source trail, last year’s questions answered, no gap blocking release, no policyholder names or street addresses. Renata signs, and version 1.0 goes to Callum Reyes for the panel of seven: a 64-page PDF, exposure data by ZIP and grid, cat events and large losses, the EP curves and own-view note, and the answers letter — 17 of 18 answered, the 18th (roof age for 31% of North Carolina homeowners) declared with its capture plan, complete by 31 Mar 2027.
“Released to Callister Re Brokers today, 11 days before the deadline. Version 1.0 is frozen; changes after this are a new version, sent with a change note.”
The same pack, seen by the five people who carry it to the market — what their October looked like, and what it looks like now.
Six agents build each chapter, tie every figure and answer last year's questions; the head of ceded reinsurance signs.
Builds the exposure chapter from the in-force policy extract and program bordereaux: TIV and locations by state, construction, occupancy, distance to coast and geocoding quality, compared with last year, with every change over 2% explained.
Pulls cat events and large losses from the claims system, trends them to the renewal year on the actuary’s approved on-level factors and applies each event as-if to the program.
Checks the modelled exposure against the exposure profile, reads the occurrence and aggregate EP curves for the vendor model and the own view, builds the bridge from last year’s 1-in-100 and reports where the program top is reached.
Writes the underwriting changes chapter from rate filings, underwriting notes and non-renewal lists — rate, appetite, deductibles and programs — each item cited.
Reads the broker’s question log from the last renewal, finds the section of the new pack that answers each question, drafts the answer with its figure and places a flag in that section’s margin.
Ties every figure across the chapters, the cover letter and the finance plan, lists the gaps with owners and due dates, and keeps the lineage of every figure.
Signs and releases to the broker. The agents propose; a named person decides.
Ask about any figure, chapter or last year’s question in plain words — or say what to add. Every answer points to the figure and the section it comes from.
Which of last year’s questions are still open?
3 of the 18 are not fully answered yet: Q2 from Aldermoor Re — roof age is unknown for 31% of North Carolina homeowners, to be declared with the plan; Q3 from Corvane Re — the Saltmarsh locations are not in the modelled exposure; Q10 from Corvane Re — two fires are missing from the large-loss list. The other 15 are answered in the pack, each with its figure.
Why is the 1-in-100 up on last year?
It rose from $468M to $512M as modelled: exposure growth adds $39M, the Cape Cod exit takes off $8M, the vendor model update adds $6M and Harborline’s own view $7M. With the Saltmarsh locations it is now $531M. The program top is reached at 1-in-148, against 1-in-226 last year.
Where does the $86.4B TIV come from?
From the policy administration system: 412,860 locations in force at 30 Sep 2026, extracted on 13 Oct. The exposure profile builder sums building, other structures, contents and loss-of-use limits by location. Last year’s pack said $78.9B.
Add the two missing large losses and tell Corvane Re where
Done. Both are in the claims system above the $5M threshold: the warehouse fire in Linden, NJ, $7.8M (claim CL-24-031877), and the marina plaza fire in Bayport, NY, $5.6M (claim CL-25-047209). The large-loss list in § 3.4 now has 12 losses, and Corvane Re’s question Q10 is marked answered there, with a flag in the margin.
13 screens from the working solution, on its sample data. Pick one to see it large.
The renewal, the panel and the rules every chapter is checked against are settings. Changes are versioned and on the audit trail.
| Setting | Default | Choose from |
|---|---|---|
| Broker and due date | Fri 30 Oct 2026 | Fri 30 Oct · Mon 26 Oct · Fri 6 Nov 2026 |
| Questions carried forward | On | Every question the panel asked at the last renewal answered in chapter 1 |
| Cross-chapter tolerance | 0.5% | Set up or down; a larger difference is a gap |
| Large-loss threshold | $5M per risk | $2.5M · $5M · $10M |
| Years of loss history | 10 years | 5 · 10 · 15 years |
| Trend factors | Megan Tsai · approved 2 Oct 2026 | The pricing actuary’s on-level factors — never set by an agent |
| Cat modelling chapter signed by | Dr. Hugo Lindqvist | The catastrophe modelling lead |
| Release to the broker signed by | Renata Solís | The Head of ceded reinsurance |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample renewal data · “target” = the design goal, measured in the live solution · “estimated” = our estimate · “typical” = published figures (Deloitte, ceded reinsurance survey — 71% of carriers say late data drives manual work in ceded reinsurance; 57% have no single source for the data they need). People, companies and reinsurers named on this page are fictional — characters and sample data in the working solution.
Software that builds the submission a cedant sends to the reinsurance market at renewal — program structure, exposure profile, loss history, cat model output and underwriting changes — from the carrier’s own systems. Reinsurance Submission Preparation adds six agents that build each chapter, tie every figure across chapters and to its source, and answer last year’s reinsurer questions in the new pack.
The question answerer reads the broker’s question log from the last renewal, finds the section of the new pack that answers each question and drafts the answer with its figure, flagged in that section’s margin. In the working solution 15 of 18 were answered when the pack was assembled; partial or open answers are declared with an owner and a plan.
No. The cat model reconciler reads the EP curves and the model input, checks the modelled exposure against the exposure profile and explains the change from last year. When the modelled exposure is short, it asks the catastrophe modelling lead to re-run, and the lead signs the chapter.
The consistency and lineage checker ties every figure across the chapters, the cover letter and the finance plan. Any difference above the tolerance — 0.5% by default — becomes a gap with an owner and a due date. It never picks one of the two numbers; the owner resolves it.
The pricing actuary. The loss history builder trends events to the renewal year only on the actuary’s approved on-level factors, applies them as-if to the program and explains development since last year’s pack. It never sets a trend factor or changes a reserve.
No. No policyholder names or street addresses leave in the pack; locations are redacted to ZIP and grid. That is one of the checks that must pass before release.
Yes. Agents build and check; people decide. The catastrophe modelling lead signs the cat modelling chapter, the Head of ceded reinsurance signs and releases the pack, and nothing is sent to the broker until she signs. Every step is on the activity trail.
The Agentic Solution Engine builds and deploys it from your requirements — your source extracts, last year’s submission and the broker’s question log — and it goes live once every quality gate has passed. We will walk you through it on your own renewal first.
A first view of every program in hours, not days.
1-min demoTreaty Wording ComparisonEvery changed clause found before signing, not when a claim tests it.
1-min demoReinsurance Contract SearchCoverage questions answered with the clause, without waiting on one specialist.
1-min demoWe’ll run Reinsurance Submission Preparation on your own source files and last year’s submission.