Case studies · Hospitality & travel
Invoice-to-booking reconciliation for a travel retailer
Around 500 bookings a day produced 1,000 to 2,000 supplier documents, each checked against the booking by hand. An AI extraction and matching pipeline now surfaces only the discrepancies.
- Client
- A cruise and package-holiday retailer
- Built with
- Azure Data Lake, Azure Data Factory, Microsoft Fabric, Azure OpenAI, Power BI
- Timeline
- Eight months from discovery to phased rollout
The problem
Supplier invoices, tickets and certificates arrived in every format the travel industry uses, and a team verified each one against the booking manually. Errors slipped through, revenue leaked when supplier invoices did not match bookings, and amendments made by customers after a booking were often not passed to suppliers.
What we built
- 1A central data platform on Azure with pipelines from the booking system and the CRM
- 2AI document extraction with a model trained on travel invoices and tickets, capturing passengers, itineraries, flights and pricing
- 3Matching rules comparing extracted data with booking records and flagging discrepancies
- 4Amendment tracking to confirm every booking change reached the supplier
- 5Dashboards for reconciliation status and exceptions
What changed
- —Reconciliation moved from reading every document to reviewing a queue of exceptions
- —Amendment gaps between bookings and suppliers became visible instead of discovered by the customer
This project became a ready-made offer you can start from.
See the offer