Upload the bordereaux, validate them against the binder, export them to the system that needs them: both do that, and both prepare SOV location data for catastrophe models. Brisc goes one step further on the same file and reconciles the premium to the cash: the bank, the remittance and the policy, with evidence on every match.
A London data cleaning platform with two products on one import, enhance and export flow. Its Delegated Authority Platform cleans, standardises and validates risk, premium and claims bordereaux: you upload the file and enter the binder-level information, it checks every row against binder terms and limits, remembers each mapping, flags column changes for review, and exports v5.2 or a custom format. Its Underwriting Platform prepares SOVs and exposure data for catastrophe models. In November 2026 Gallagher Re announced it uses Scrub AI for property binder bordereaux and exposure data. Scrub AI names PDF slip ingestion, email ingestion and ScrubHub as what comes next.
The Bordereaux Analyst takes risk and premium, claims and cession bordereaux in each partner’s layout, maps each layout once, standardises the fields, removes duplicates, validates every row against the binder and the last version as the file lands, and exports the bordereau in the layout your system or the receiver loads, including v5.2. The Reconciliation Analyst then matches the premium on it to bank statements and remittance advices, at policy level, and builds the capacity bordereau from matched cash. The Submissions Analyst handles the SOVs that come with a submission: it builds a per-location schedule, geocodes the locations the documents don’t place, codes occupancy and stated construction to your template’s codes, and fills the pricing and modelling templates your downstream systems load. Every check is a named rule you can read.
| Scrub AI | Brisc | |
|---|---|---|
| The job | Clean, standardise and validate SOVs and bordereaux so the data is ready to use | Clean, standardise and validate bordereaux and export them; then reconcile the premium to the bank, the remittance and the policy |
| Bordereaux in | Risk, premium and claims bordereaux, uploaded with the binder-level information entered by the user | Risk and premium, claims and cession bordereaux in each partner’s layout, mostly Excel and CSV, by upload, email or a scheduled pull |
| Cleaning | Cleans and standardises every row; missing or incorrect mandatory fields flagged | Header rows found and sheets combined; dates, amounts, codes, claim status and broker names standardised; duplicates and blank rows removed; every version kept |
| Layout changes | Remembers each mapping and flags monthly or quarterly column changes for review | Each layout mapped once and reused. Safe column changes are applied and noted; anything else pauses the file with the reason in plain English |
| Validation | Every row against binder terms and limits; missing or incorrect mandatory fields flagged | 50+ checks on every row as the file lands: commission rate against the contract, tax and levy, policy validity, currency, footing, duplicates, changes since the last accepted version and rows that have gone missing |
| Claims bordereaux | Cleaned and validated alongside risk and premium | Versions compared, with reserve movements, new large losses, reopened claims and watch-list keywords flagged and the before and after shown |
| Output | v5.2 or a custom format; API into existing workflows and binder management tools | The bordereau in the layout your system or the receiver loads, including the Lloyd’s Coverholder Reporting Standards v5.2 file, as a file or by API |
| The cash | Not described on its product pages | Reported premium matched to bank statements and remittance advices, and the capacity bordereau built from matched cash. See bordereaux reconciliation |
| SOVs and cat models | Address validation and geolocation, occupancy and construction coding, TIV flags, export to RMS, AIR, OED and Sequel Impact | The Submissions Analyst builds the per-location schedule, geocodes the locations the documents don’t place, codes occupancy and stated construction to your template’s codes, and fills the location template your modelling platform loads, such as Moody’s RMS |
| How it works | Built, in Scrub AI’s words, on “deterministic machine learning models” trained on insurance data | Every check and every match is code against a named rule you can read. AI reads the documents and suggests new mappings, which a person confirms |
| Exceptions | Issues flagged for a person to review | Each flag queued with its reason and source document; nothing is corrected without a person deciding |
| Time to live | Not stated | 2 to 6 weeks |
Scrub AI column drawn from its own pages, as read on 7 November 2026: scrub-ai.com, its Delegated Authority Platform page, its Underwriting Platform page, its Gallagher Re release of 2 November 2026 and its podcast recap of 19 October 2026. How the layers fit together: bordereaux management, who does what.
When clean data is the whole deliverable: bordereaux and SOVs standardised for the tools you already run, with the cash reconciled somewhere else.
When you want the bordereau cleaned, validated against the binder and exported, and then the answer to whether the premium on it was paid. Scrub AI stops at the clean, validated file. Brisc does that part too, prepares the SOVs as well, then takes the same file on to the bank, the remittance and the policy, with the rule and the evidence behind every match and every exception.
Bring a month of bordereaux with the bank statements and remittances that should settle them. Compare what each tool flags, and what you can prove at the end.
Related: Brisc vs VIPR · VIPR alternatives · Bordereaux management software: a buyer’s guide · Bordereaux Analyst · Bordereaux reconciliation · Reconciliation Analyst
On bordereaux, both do the same first job: take the file in the coverholder’s layout, clean and standardise it, validate every row against the binder, and export it in the format the receiving system loads, including v5.2. Both also prepare SOV location data for catastrophe models. Brisc then does the step Scrub AI does not describe: the Reconciliation Analyst matches the premium on the validated bordereau to the bank, the remittance and the policy, with evidence on every match.
Yes. The Bordereaux Analyst takes risk and premium, claims and cession bordereaux by upload, email or a scheduled pull, mostly as Excel and CSV. It maps each partner layout once, standardises dates, amounts, codes, claim status and broker names, removes duplicates and blank rows, and keeps every version. It runs 50+ checks on every row as the file lands, including commission rate against the contract, tax and levy, policy validity, currency, footing and changes since the last accepted version. Then it exports the bordereau in the layout your system or the receiver loads, including the Lloyd’s v5.2 file.
Yes. The Submissions Analyst builds a per-location schedule from the SOV, slip or inspection report, geocodes the locations the documents don’t place and labels where every coordinate came from, codes occupancy and stated construction to your template’s own codes, and fills the location template your modelling platform loads, such as a Moody’s RMS template, alongside your pricing template.
Its product pages, as read on 7 November 2026, describe cleaning, standardising and validating bordereaux against binder terms and limits, then exporting them. They do not describe matching bordereaux to bank statements or remittance advices. That is the step Brisc adds: a validated bordereau becomes the expected side of the cash match.
Yes. A file Scrub AI has already cleaned is a good input: Brisc reads it like any other partner file, validates it against the binder and the last version, and reconciles the premium on it to the money that arrived. Brisc also does the cleaning, validation and SOV preparation itself, so one tool can cover the file, the exposure data and the cash.
A named rule produced the number, and you can read the rule. Every bordereaux check and every cash match in Brisc is code against a named rule, so a flag is reproducible and a total is computed, never generated. AI reads the documents and suggests how a new partner layout maps to your schema, which a person confirms. It never decides a match.
Live in 2 to 6 weeks. In production the Bordereaux Analyst has processed, mapped, de-duplicated and checked 360,000+ bordereau rows and flagged 23,000+ discrepancies for human review. On the cash step, expect about 80% of records matched on day one, rising to 92 to 95% by about 90 days as the rules absorb your book.
30 minutes, no slide deck. Tell us where the volume hurts and we show you the Analyst on it. Bring a bordereau, statement or submission if you have one; you don’t need to.
Book the walkthroughNot ready for a call? Get the Decoding AI guide →