AI in travel: Finding patterns in BSP exceptions

Why BSP reconciliation exceptions take time

For a travel agency selling airline tickets, BSP reconciliation is a regular part of financial control. The finance team compares transactions reported through the Billing and Settlement Plan (BSP) with the tickets and amounts recorded in its own systems. Many entries can be matched automatically and require little effort, but the remaining differences require someone to work out what happened and decide how to resolve them… and this is where BSP reconciliation becomes time-consuming.

A commission variance is a good example. The BSP data amount differs from the amount the agency posted for a ticket, but the difference alone does not explain why. A finance professional may need to check the ticket, the airline, the applicable commission agreement, and any related adjustments. When several tickets show a variance, reviewing them one by one can make it difficult to see whether they share an underlying issue.

Other exceptions also require investigation, including missing tickets, unexpected amounts, refunds, and airline debit or credit memos. The team needs enough context to determine whether the agency’s records should change, the BSP information needs to be questioned, or the difference is expected.

From a list of differences to a useful explanation

Rules-based matching helps agencies reconcile transactions that meet defined conditions. The time-consuming work often sits in the exceptions left over, particularly when the information available does not immediately explain why an amount differs.

Reviewing those exceptions together can reveal patterns that are easy to miss when tickets are reviewed individually. For example, a similar commission variance across several tickets might prompt the team to check whether the same agreement applies to all of them or whether commission was recorded incorrectly. The investigation can then start with a possible shared cause instead of beginning again with each ticket.

An AI agent can support this work by analyzing a group of exceptions and presenting patterns for the finance team to assess. Its findings help guide the investigation, especially when a recurring issue affects multiple transactions.

How the TravelOperations BSP reconciliation agent supports finance teams

The TravelOperations BSP reconciliation agent supports this investigation. Once the finance team has loaded a BSP file and reconciled tickets that meet the matching rules, it can use the agent to analyze any remaining commission differences.

The agent examines the discrepancies and presents its findings in plain language, along with a suggested cause, a recommendation, and an indication of uncertainty. If several tickets for the same airline show a similar variance, the findings may point the team toward a commission agreement that needs checking. Finance professionals can use that lead to review the relevant records and decide how to handle the discrepancies in the reconciliation workspace.

Before acting on a suggested pattern

A shared pattern can focus the investigation, but the team still needs to confirm that the suggested cause applies to the transactions involved. Before resolving several exceptions together, it may need to check:

• whether the tickets fall under the same airline commission terms
• whether a refund, change, or adjustment explains an individual variance
• which amount is supported by the agreement and the agency’s records
• whether every ticket calls for the same resolution

These checks may confirm a recurring commission issue or reveal a ticket that needs separate treatment. They also help the team decide whether the issue extends beyond the tickets currently under review.

See the BSP reconciliation agent in action in our on-demand webinar.

A practical place to start with AI in travel finance

BSP reconciliation gives travel agencies a focused way to evaluate AI because the finance team already understands the process and its recurring exceptions. An agency could start with one type of discrepancy, see whether the agent identifies useful patterns, and compare its suggestions with its specialists’ conclusions.

That comparison helps the team judge where the agent adds value and what information it needs to investigate exceptions reliably. Over time, it may also reveal recurring problems in commission data or processes worth addressing at the source.

---

Other articles in this series

AI in travel: From manual reconciliation to intelligent exception management