Natural Language Processing
What if processing a supplier quotation took minutes instead of half an hour? At CFL, AI is taking over the repetitive work behind quotation processing, turning unstructured supplier documents into ready-to-use data in SAP. The result is a faster procurement process and more time for buyers to focus on what actually requires their expertise.
Supplier quotations arrive in different formats, often by email or PDF. For Chemins de Fer Luxembourgeois (CFL), buyers were spending up to 35 minutes per quotation reading, interpreting and entering data manually. That left less time for analysing offers, working with suppliers and making sourcing decisions.
Together with Flexso, CFL introduced SAP Document AI on SAP Business Technology Platform to automate the process. The AI reads and interprets supplier quotations, even when they come in different and unstructured PDF formats. The extracted data is then validated and transferred into SAP.
By connecting the solution directly to the mailbox, the process became fully hands-off for buyers. They no longer need to interact with the AI themselves. Instead, AI handles the repetitive data work in the background, while buyers stay focused on decisions and supplier relationships.
1
70% faster quotation processing, reducing processing time from 35 to 10 minutes.
2
Around 30 days of efficiency gains per year.
3
A scalable first step into AI, with strong adoption and a foundation for further automation across procurement.
The results were immediate. Quotation processing went from around 35 minutes to 10 minutes, making the sourcing process up to 70% faster. That saves around 25 minutes per purchase order, adding up to approximately 30 days of efficiency gains each year.
The AI reached an 80% recognition rate across different quotation formats after seven weeks. Adoption also grew quickly, withmore than 300 quotations created by AI in the first seven weeks and 60% user adoption within the first month.
This project was CFL's first productive AI use case. Beyond improving one procurement process, it gave the organisation apractical foundation for further AI and automation across procurement and beyond.
For buyers, the impact is simple: less time spent on data entry, fewer manual errors and more time for the work that needs human expertise.
The initiative was also nominated for the SAP Innovation Awards 2026 in the AI Excellence category.
.jpg)
.jpg)
%20copy.webp)
