Monday morning. Your credit controller opens the HSBC statement and finds a payment from a European broker that is £8,400 less than the bordereau says it should be. The bordereau quotes gross written premium. The bank receipt is net of UK Insurance Premium Tax at one rate, broker commission on a sliding scale that resets quarterly, and a German Versicherungsteuer adjustment that applies to the property sub-line but not casualty. Three deductions, applied by three different counterparties, at three different rates — governed by a tax schedule that changed in January for a policy that incepted in December.
She has twelve minutes before the morning allocation meeting. She knows the answer is in a combination of the IPT schedule, the commission table, and last quarter's rate change notice from the broker. She also knows that the rate change notice is in an email attachment from three months ago, and the commission table is in a spreadsheet that her colleague — who left last month — maintained by hand. It's the same "Excel archaeology" every credit-control desk runs — the subject of Excel Archaeology: A Day in the Life of an Insurance Cash Matcher.
Brisc AI is an insurance-native AI platform that automates cash matching and bordereaux reconciliation for MGAs and reinsurers. We built the Reconciliation Analyst because of scenes like this one — because every operations leader we talk to describes the same structural problem. Cash matching accuracy depends entirely on knowing what "correct" looks like before you can attempt a match. And in a multi-jurisdiction operation, "correct" is a moving target — one piece of the larger Cash Ops function most insurers have never formally staffed for.
Cross-jurisdiction tax variance is the structural reason that insurance cash matching is harder than it looks from the outside. When a policy spans multiple jurisdictions — or when an MGA writes across several — the gross premium on a bordereau and the net amount arriving at the bank are never the same number. The gap between them is filled by a stack of deductions that varies by jurisdiction, by line of business, and by the effective date of the policy.
The three layers of that stack:
Jurisdiction-level taxes. UK Insurance Premium Tax applies at 12% on most general insurance, but 20% on travel and some appliance warranties. German Versicherungsteuer runs at 19% on most classes but 22% on fire insurance. EU stamp duties vary by member state, with France applying a fixed-percentage tax on top of premium. US surplus lines taxes range from 1.5% to 6% depending on the state, with additional stamping fees in some jurisdictions. Each of these taxes is calculated differently — some on gross premium, some on net, some on the premium after other deductions have already been applied.
Line-of-business variance. Within a single multi-line program, property and casualty often carry different tax treatments in the same jurisdiction. A combined property-casualty bordereau from a Lloyd's coverholder can contain policies where IPT is calculated on different bases for different sub-lines within the same remittance. The credit controller cannot match the aggregate payment without decomposing it into its sub-line components — and the bordereau does not always make that decomposition explicit.
Temporal complexity. Tax rates change. UK IPT moved from 6% to 10% in 2015 and to 12% in 2017. When a rate changes mid-policy, the bordereau may reflect the old rate while the remittance reflects the new one — or vice versa, depending on whether the broker calculates tax at inception or at settlement. The credit controller is now matching against an expected amount that depends on which date the broker used, which the bordereau does not specify.
One MGA we work with operates across six geographies with thirty-eight binder partners — the same anonymized shape behind MGAs: When Cash Matching Becomes Your Growth Constraint. Each partner's remittance follows its own convention for netting taxes and commissions. McKinsey and Accenture estimate that 30-40% of operations time in insurance goes to administrative tasks. At this MGA, a disproportionate share of that time lives in one place: re-deriving what "correct" means before every match attempt, for every jurisdiction, for every line.
Deterministic matching works when the expected amount is static. If the rule says "gross minus 12% IPT minus 3% brokerage," it clears the straightforward cases — exact amounts, exact references, exact timing. Those cases might represent half the book on a good quarter.
The other half is where it breaks. The part-payment where a broker has netted two programs into a single remittance. The over-payment where a cedant prepaid three months of premium at a rate that was correct when the payment was sent but incorrect when it arrived because a mid-year IPT adjustment changed the basis. The payment that appears £1,200 short because the broker applied a sliding-scale commission tier that the MGA's finance system still has at the previous tier.
These cases fall back to people. And the people who know the answer — who remember that Broker X nets German Versicherungsteuer before remitting but Broker Y sends gross and expects the MGA to self-assess — are the same experienced analysts whose institutional knowledge walks out the door when they leave. Industry data points to 20-40% annual turnover in insurance back-office roles, with a 90-to-180-day ramp before a new hire becomes productive in cash matching. Every departure resets the clock on jurisdiction-specific knowledge that took months to accumulate.
A $47,000 discrepancy at one MGA took three months of broker correspondence to resolve. The money was there — correctly received, correctly deposited, incorrectly labeled. The label was wrong because the commission calculation assumed a flat rate when the contract specified a sliding scale with a quarterly reset. The analyst who knew the contract terms had left six weeks earlier.
The gap is not computational power. Calculating UK IPT at 12% is trivial. The gap is remembering that this broker, for this program, in this jurisdiction, applies tax at this point in the chain, nets commission before remitting, and resets the sliding scale in March rather than January — and retaining that knowledge permanently, across every cedant, every jurisdiction, every tax-rate change.
Brisc's Reconciliation Analyst accumulates every deduction pattern it encounters into a permanent dictionary. The first time a broker remits a German property premium net of Versicherungsteuer at 19% alongside a UK casualty premium net of IPT at 12%, with commission netted at different tiers for each sub-line, the system learns the pattern. The second time that broker sends a similar remittance, it matches automatically — not because a rule was written, but because the institutional knowledge was retained.
Helix Underwriting Partners reports that Brisc removed 80% of manual labour from their operations. The mechanism is knowledge retention — not faster calculation, but permanent institutional memory of every deduction structure, every commission convention, every jurisdiction-specific tax treatment across every cedant the system has processed.
Brisc's Reconciliation Analyst operates at 97%+ accuracy on bordereaux reconciliation, and the accuracy does not degrade when programs span multiple jurisdictions or when tax rates change mid-year. When UK IPT moved from one rate to another, a rules-based system needs a human to update the rate table and retroactively correct the affected matches. An insurance-native system that has seen the transition pattern across its entire history of matches recognizes the shift in the data itself.
The 59% labour cost reduction that Brisc customers report is not a speed improvement. It is the elimination of the re-derivation loop — the analyst no longer spends thirty minutes re-building the expected-amount calculation for every match attempt in every jurisdiction. The system already knows what "correct" looks like, because it remembered the last time it looked.
Typical deployment takes 2-6 weeks, and the system achieves 80% auto-match accuracy on day one — before the jurisdiction-specific knowledge begins compounding. By day ninety, the accumulated dictionary covers the long tail of broker-specific conventions that no rules table could anticipate.
What is cross-jurisdiction tax variance in insurance?
Cross-jurisdiction tax variance is the difference between gross written premium recorded on a bordereau and the net amount that arrives at the bank, caused by jurisdiction-specific taxes, duties, and deductions. UK Insurance Premium Tax, German Versicherungsteuer, EU stamp duties, and US surplus lines taxes all create different gaps between billed and received amounts — making automatic cash matching difficult for any MGA or reinsurer operating across multiple geographies.
Why does UK Insurance Premium Tax make cash matching harder?
UK IPT applies at 12% on most general insurance but 20% on travel insurance and some appliance warranties. Within a single multi-line program, different sub-lines may carry different IPT rates. Brokers handle IPT differently — some remit net of IPT, others remit gross and leave the MGA to self-assess. These inconsistencies mean the expected amount varies by broker, by line, and by convention, not just by rate.
How do EU stamp duties affect bordereaux reconciliation?
EU member states apply different stamp duties and premium taxes at different rates, calculated on different bases. France applies a fixed-percentage tax on premium. Germany levies Versicherungsteuer at 19% on most classes but 22% on fire insurance. A single bordereau covering policies across three EU jurisdictions can contain three different tax treatments — and the remittance arrives as one aggregated payment. Decomposing that payment into its jurisdiction-level components is where manual reconciliation spends its time.
What are US surplus lines taxes and how do they create matching problems?
US surplus lines taxes range from 1.5% to 6% depending on the state, with additional stamping fees in jurisdictions like Florida and Texas. For an MGA writing excess and surplus lines across multiple states, a single broker remittance may net taxes at different rates for different insured locations within the same program. The bordereau typically records policies at gross premium; the bank statement reflects the net amount after state-specific deductions.
Can rules-based reconciliation handle cross-jurisdiction tax variance?
Rules-based matching handles static cases — exact amounts at known tax rates. It does not handle the long tail: mid-year rate changes, broker-specific netting conventions, sliding-scale commissions with quarterly resets, or aggregate payments that combine multiple jurisdictions into one remittance. These cases require institutional knowledge of how each broker, in each jurisdiction, actually calculates and remits — knowledge that rules tables cannot capture and that experienced analysts take with them when they leave.
How does Brisc's Reconciliation Analyst handle tax-variant payments?
Brisc's Reconciliation Analyst accumulates every deduction pattern it encounters — every broker convention, every tax-rate variation, every commission structure — into a permanent dictionary. Rather than applying a static rule table, it matches incoming payments against learned patterns from prior cycles. When a broker's remittance convention changes or a tax rate shifts, the system recognizes the new pattern in the data without requiring a manual rule update. The result is 97%+ accuracy across multi-jurisdiction books.
What happens when tax rates change mid-policy?
Tax-rate changes create a temporal matching problem: the bordereau may reflect the old rate while the remittance reflects the new one, or the reverse. Brisc's Reconciliation Analyst handles this by retaining the full history of tax treatments for each broker and program. When it encounters a payment that matches the new rate but the bordereau still shows the old rate, it recognizes the transition pattern rather than flagging a false exception.
How long does it take to deploy Brisc across multiple jurisdictions?
Typical deployment takes 2-6 weeks, depending on the number of programs and the complexity of existing broker formats. The system achieves 80% auto-match accuracy on day one by applying its accumulated dictionary of insurance-specific patterns, and improves to 92-95% over the first ninety days as it learns jurisdiction-specific conventions unique to each client's broker network.
If your credit control team spends more time calculating what a payment should be than matching what it is, you're looking at a tax-variance problem that compounds with every new jurisdiction and every new program. See how Brisc's Reconciliation Analyst handles your data →