Case Study
How Unilight made commission recovery faster with exception-based reconciliation
Unrecovered brokerage is rarely visible as one obvious mistake. It is usually hidden across insurer statements, policy records, invoices and follow-up notes.
At Unilight Reinsurance Brokers, a seven-person team was spending weeks matching insurer statements and following up on differences. The work was important, but too much effort went into checking transactions that were already correct.
The difficulty with line-by-line reconciliation
In a manual process, every entry looks like work. The team compares policy numbers, premium, expected brokerage and the amount received—often across files with different formats and references.
As volume grows:
- A missing or differently formatted policy number can prevent a match.
- A short receipt can remain hidden inside a large statement.
- Employees may repeat checks because there is no shared verification history.
- Genuine collection issues compete for attention with correctly paid transactions.
The result is more than slow reconciliation. Skilled employees spend their time proving that normal transactions are correct instead of recovering what is genuinely pending.
Changing the unit of work
With SIBRO, reconciliation became exception-based.
Rather than presenting the entire insurer statement as a task list, the process separates matched records from items that require attention. The team can then focus on exceptions such as:
- Expected brokerage not received
- Short or excess receipts
- Transactions without a matching policy
- Policy or invoice reference differences
- Amounts that require a rule or data correction
The underlying policy, premium and financial context remains available, so an exception can be investigated rather than merely listed.
Before and after
Before
- Seven people involved in reconciliation activity
- Weeks spent matching statements and following up
- Correct and incorrect transactions reviewed together
- More effort spent searching than resolving
After introducing exception-based reconciliation
- The activity could be handled by three people
- The team concentrated on genuine differences
- Commission recovery became faster
- The cost of recovering pending commission reduced
Automation created value by removing already-matched transactions from the work queue—not by removing the judgement required to resolve an exception.
A better exception queue asks better questions
For every difference, the team should be able to ask:
- Was the expected brokerage actually received?
- Does the insurer’s reference match the policy or invoice in another format?
- Is the difference caused by booking data, a pay-in rule or an actual short payment?
- Who owns the insurer follow-up?
- What was the last action, and what remains pending?
This turns reconciliation from a periodic spreadsheet exercise into an operational control.
What other broking firms can learn
Reconcile continuously. A difference is easier to understand when it is found close to the transaction date.
Work from exceptions. A long insurer statement is not a useful task list. A shorter, prioritised list of differences is.
Connect accounts to operations. Brokerage and receipts are easier to investigate when the underlying policy, premium, branch and owner are visible.
Use recurring errors to improve the process. Reconciliation can expose incorrect rules, incomplete booking, missing references and training gaps—not only payment gaps.
Keeping the result realistic
Every implementation is different. Results depend on transaction volume, statement quality, available references and the discipline with which the team follows the process. The Unilight outcome should therefore be understood in its own operational context.
The broader principle remains useful: when software removes correctly matched work from the queue, people can concentrate on recovering what is genuinely pending.