Reinsurance contract search that answers coverage questions with the clause that decides
Coverage questions answered with the clause, without waiting on one specialist.
See one case, screen by screen ↓Software that answers coverage questions about your reinsurance program from the treaty wordings themselves. Reinsurance Contract Search reads every wording, slip and endorsement article by article, and a plain question returns a short answer with the deciding clause quoted, plus the same question answered across every treaty in the program in one table.
How to choose AI solutions for insurance →
Updated
When a large loss lands, claims, finance and underwriting all ask the same thing: does our reinsurance respond, under which clause, and for how much? The answer sits in five treaties a year — quota share, per risk, catastrophe, aggregate, casualty — each with its own wording, slip, endorsements and structure chart. Only one or two people on the ceded team know them well, so the question joins their queue behind bordereaux and renewals.
And the answer has to come from the wording in force on the loss date, endorsements applied, read in inuring order. Miss one renewal change and the answer is wrong with confidence. Nobody reads the same clause across all five treaties until a recovery is refused — or a notice deadline passes and a reinsurer has grounds to decline.
Estimated split for one large-loss coverage question, answered by email from the specialist and with the solution.
Seven specialist agents read the program and quote the clause that decides; the ceded analyst checks and the Head of Ceded Reinsurance signs.
A ransomware attack locks the refrigeration controls at Kettleby Cold Storage in Joliet, Illinois, and an explosion and fire follow. Here is how the question “does our reinsurance respond?” is answered, screen by screen, in the working solution.
Tom owns claim CM-26-04417: loss 3 Oct 2026 at 02:40, gross $14.6M. He asks “Does our reinsurance respond to the Kettleby fire caused by the ransomware attack?” It lands in the ceded team’s Waiting for you list.
“Wordings in force on the loss date · Every sentence carries its clause”
The loss fact reader takes the cause the adjuster recorded — ransomware on the refrigeration controls — and the amounts. The clause finder picks the wordings in force on 3 Oct 2026: five treaties, four endorsements applied, 2025 wordings kept for comparison. It searches 3,912 clauses; 14 articles match and 5 decide. The response assessor applies the inuring order, quota share first.
“Citation checker · 4 of 4 sentences carry a clause — one sentence rewritten to quote Art. 9.3 word for word.”
The quota share responds for 30%, $4.38M, under Art. 12.3. The per risk treaty does not respond under Art. 9.2–9.3, so $8.0M stays net. The catastrophe treaty is not reached — one risk, $10.22M net, below the $50M retention. The aggregate excludes it, and the casualty treaty responds only above its $5M retention, with 31 claimants reserved at $1.9M.
“Partly. $4.38M is recoverable; $8.0M the per risk treaty would have paid stays net.”
Nadia opens the Property Per Risk Excess 2026 at Art. 9 Cyber Loss. The deciding words are highlighted: the fire and explosion write-back applies “unless that Cyber Incident is caused by … a Cyber Act”, and Art. 9.4(a) defines a Cyber Act as an unauthorised, malicious or criminal act involving a computer system. The margin shows what the clause decides, its cross-references and the 2025 wording beside it.
“Renewal slip line 14, agreed 18 Dec 2025, removed the Cyber Act write-back.”
One switch answers the same loss under the 2025 wordings. Both property treaties would have responded: $12.38M recoverable, because the 2025 Art. 9.3 gave back fire from “a Cyber Act or a Cyber Incident”. The answer says plainly that this is a what-if — the loss happened in 2026, so the 2026 wording decides.
“The 2026 renewal removed the words ‘a Cyber Act or’, which is the whole $8.0M difference.”
Clauses compared puts the cyber clause of every treaty side by side, tested on a ransomware fire like Kettleby. The quota share follows the Kettleby form; the per risk was narrowed at the 2026 renewal; the aggregate excludes cyber loss without exception; the casualty treaty gives back third-party injury. Nadia flags the difference for the 1 Jan 2027 renewal, and Simone Ricard and Gareth Olufemi are told.
“Two treaties follow a malicious-act loss, two do not, one partly — a ransomware fire is covered on the original policy and only partly reinsured.”
Nadia checks the answer and saves it as coverage note CN-26-031, drafted by the note writer: the question, then four numbered sentences, each citing its clause — including the notice owed to the per risk and quota share reinsurers by 2 November 2026 and the quota share’s $4.38M that may be cash-called. The record shows who did what, minute by minute.
Nadia sends the note to Gareth. He signs as Head of Ceded Reinsurance, with the position stated: quota share responds $4.38M, per risk does not respond. The note goes to Tom Akers and Mei-Lin Tan with the eight clauses it cites.
“Signed · Tom Akers sets the net reserve, Mei-Lin Tan books $4.38M recoverable.”
The question log shows who asked what, from which team, and what came back: 186 questions this quarter, 152 answered without the specialist, a median of 9 s to a cited answer, and 34 referred. A question no clause answers with confidence — like whether the aggregate covers wildfire smoke damage with no serial number — waits for a specialist to take it.
“Referral: confidence below medium goes to the specialist.”
The same loss, seen by the five people who carry it — the ceded team, claims, finance and underwriting — and what their week looked like before.
Seven specialist agents read the program and quote the clause that decides; the ceded analyst checks and the Head of Ceded Reinsurance signs.
Reads every treaty wording, slip and endorsement, splits it into articles, and links each endorsement to the article it amends with its effective date.
Builds the program from the signed slips and the structure chart: layers, retentions, limits, reinstatements, inuring order and each reinsurer’s share.
Reads the claim file for the facts a clause turns on: loss date and time, peril, cause as recorded by the adjuster, location, line of business and amounts.
Finds, in each treaty, the article that decides the question in the version in force on the loss date, endorsements applied, and the related articles — with last year’s wording when a renewal changed it.
Applies the treaties in inuring order and states for each one whether it responds, does not respond, is not reached or responds above its retention — with the deciding words, the amount and the arithmetic.
Reads each sentence of the answer back against the quoted clause, rewrites a sentence to quote the clause where the paraphrase drifts, and removes anything the clause does not say.
Drafts the coverage note from a checked answer: the question, then one numbered sentence per finding, each citing its article.
Signs every note that leaves the team. The agents propose; a named person decides.
Anyone the ceded team allows can ask in plain words. Every answer quotes its clause — and where no clause answers, the question goes to a specialist instead of getting a guess.
Which notices are owed on the Kettleby loss, and by when?
Two, both by 2 November 2026: the per risk reinsurers, because the loss is over 50% of the $2M retention — even though that treaty excludes it — and the quota share, because it is over $5M. The casualty reinsurers need notice only when the injury reserve passes $2.5M; it is $1.9M today.
Compare the cyber clauses across the 2026 program
Five treaties, four different answers. The quota share follows the original form, the per risk treaty gives back fire only if the act was not malicious, the aggregate excludes everything, and the casualty treaty gives back third-party injury.
Has a cyber fire been answered before?
Once: coverage opinion CO-2025-014, the Brightwater Foods server-room fire in March 2025. That fire was accidental, and the 2025 per risk wording gave back fire from any cyber event, so the treaty responded. Kettleby differs twice: the act was malicious, and the 2026 wording no longer gives it back.
Draft the coverage note to Tom Akers
Drafted as CN-26-031, four sentences, each citing its clause. It goes to Gareth Olufemi for signature before Tom and Mei-Lin see it — a note that leaves the ceded team is always signed.
13 screens from the working solution, on its sample data. Pick one to see it large.
How answers are given, who may see what, and who decides — all settings the ceded team owns.
| Setting | Default | Choose from |
|---|---|---|
| No sentence without a quoted clause | Locked on | Always on for every answer |
| Answer from the wording in force on the loss date | On | On · Off — endorsements applied by their effective date |
| Show last year’s wording when a renewal changed the clause | On | On · Off |
| Show amounts to underwriting | Off | On · Off — off means wording and response only |
| Refer to a specialist below | Medium confidence | Low · Medium · High |
| Property and casualty specialists | Nadia Brennan · Ravi Menon | Referred property questions · referred casualty questions |
| Coverage notes signed by | Gareth Olufemi | Any note to a reinsurer, or over $1M |
| Notice reminders | 10 days before | 1 to 30 days before a notice of loss is due |
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 program · “estimated” = our estimate from ceded team practice (2–5 days waiting on the specialist; the hours split above) · “typical” = published figures (Deloitte UK, ceded reinsurance survey — 63% of large carriers say they lack ceded reinsurance specialists). Cardale Mutual, its reinsurers, insureds and losses, and the people named on this page are fictional — characters and sample data in the working solution.
Software that answers coverage questions about your reinsurance program from the treaty wordings themselves. Reinsurance Contract Search reads every wording, slip and endorsement article by article, and a plain question returns a short answer with the deciding clause quoted, plus the same question answered across every treaty in the program in one table.
The clause finder answers from the wording in force on the loss date, with endorsements applied by their effective date. When a renewal changed the clause, last year’s wording is shown beside it — clearly as last year, or as a labelled what-if — and never used to decide the answer.
Yes. The response assessor applies the treaties in inuring order and states for each one whether it responds, does not respond, is not reached or responds above its retention, with the deciding words and the amount. In the working solution, one cyber question is answered across all five 2026 treaties in 9.1 seconds.
No. The agents never notify a reinsurer, never commit a recovery and never decide a disputed meaning. The ceded analyst checks every answer, and any coverage note to a reinsurer or over $1M is signed by the Head of Ceded Reinsurance before it leaves the team.
It says so rather than guess. If the clause finder’s confidence is below the referral setting (medium by default), the question goes to the named specialist for that line with what was searched, and it stays open in the question log until someone takes it.
Yes, each with its own view. Claims sees the program year of its losses; finance sees every program year with amounts, accounts and cash calls; underwriting sees wording and response but no amounts. Facultative certificates and counsel opinions stay with the ceded team.
Treaty and facultative wordings, slips, endorsements, program structure charts and prior coverage opinions — PDF, Word or scans. It reads loss facts from your claim system and the booked program from your reinsurance ledger, both read only.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your wordings, slips, structure charts and a sample of past coverage questions — and it goes live once every quality gate has passed. We will walk you through it on your own program first.
A checked renewal submission in days, not weeks, every figure traced.
1-min demoTreaty Underwriting WorkbenchA 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 demoWe’ll run Reinsurance Contract Search on your own program and a few of your past coverage questions.