The four numbers a broker's MD actually cares about, readable in five seconds: what we should have been paid, what we were paid, the gap this month, and how much the service has recovered for them so far this year. That last one is the sales weapon. "Recovered YTD" is the line that renews the subscription, because it makes the £400/month fee visibly self-funding.

Source: expected figures from the broker's own back-office system export (Acturis/OpenGI); received figures from the insurers' statements.

The control table. One row per insurer answers "are we square with everyone?" and shows statement receipt status, because insurers sending statements late (or not at all, like Markel here) is itself a common leak nobody tracks. A clean row builds trust in the matched numbers; a red row directs attention.

Source: every line of every insurer statement, summed and matched against every live policy in the ledger for the period.

This is the product. Each row is a specific policy where the insurer paid less than the agency agreement says they owe, with the likely cause diagnosed (wrong commission rate applied, a mid-term adjustment missing, a policy simply absent from the statement, a clawback taken in the wrong period). These are the £150–£600 discrepancies that die unnoticed in spreadsheets; across a year they're thousands. The "Action" column shows queries were already raised. The report doesn't just find problems, it evidences the chase.

Source: line-level comparison of statement vs ledger. The diagnosis logic (rate mismatch vs missing policy vs timing) is the domain knowledge the automation encodes, and the reason this service is hard to replicate with a generic tool.

The reverse check: money appearing on statements that doesn't correspond to any policy we know about. Sometimes a pleasant surprise, sometimes another broker's credit mis-posted (which will be clawed back later, so better to know now), sometimes a legacy policy the ledger lost. Its presence tells the client the reconciliation runs both directions, which is what makes it an audit-grade document rather than a summary.

Source: statement lines that failed to match any ledger record.

Cancellations and mid-term changes are where legitimate reductions live, and where illegitimate ones hide. Verifying "these 4 clawbacks are all correct" is as valuable as flagging errors: it's the section that stops the broker's finance person re-checking everything by hand out of distrust, which is the habit we're being paid to end.

Source: statements cross-checked against the ledger's cancellation and MTA records.

Memory. Underpayment chases stretch across months and die when the spreadsheet's owner gets busy or leaves. Carrying every open query forward until it resolves, with a visible "Resolved ✓" when the money lands, is what turns a monthly document into a continuous service, and it quietly compounds the Recovered-YTD number on page one.

Source: the service's own query log, rolled forward each month.

Trust and liability in one paragraph: exactly which statements and which ledger export produced the numbers, what's missing, and a plain statement that it's a reconciliation aid checked before dispatch. That is the honest disclaimer that keeps a data service on the right side of "not financial advice" while still being accountable.

Also doubles as the onboarding spec. It tells a prospect precisely what they'd need to send us: insurer statements as received, plus one monthly ledger export. Nothing else.