A global leader in supply chain solutions for the hospitality sector, engaged in sourcing, procurement, and logistics across hotels worldwide. With a presence in multiple regions, the company helps streamline operations for thousands of properties around the globe.
Handling over 75-85 million transactions annually from over 3,500 suppliers dealing with more than 550,000 products, the client orchestrates end-to-end procurement experience for 19,000 hotels worldwide. The massive proportions of their operations means that they are in a great position to successfully negotiate vendor contracts for better pricing and secure 15-40% savings for customers. However, this also presents a huge data challenge: there are multiple systems through which the data comes in, all differently organized, resulting in millions of rows of unstructured data that require processing.
Our client engaged rebate specialist, Enable, to step up their automation and bring in a streamlined incentive management program.
Enable quickly recognized that clean, reliable data was the cornerstone of automation success. And the data on hand was unstructured and unsuitable for processing in its current state.
This is where dataX.ai stepped in.
dataX.ai delivered an automated, scalable, and intelligent solution powered by our pre-trained ML models and LLM techniques, specifically designed for unstructured product data processing. These models are built to handle inputs such as supplier feeds, transactions, and invoices—by extracting attributes, validating them, and normalizing inconsistencies.
Specifically, these were the tasks that we undertook:
1. De-duplication and Product Matching with Precision:
At the time of this writing, we have processed over 30 million records, with ongoing efforts to scale further—empowering the client to continuously reduce redundancy, improve data reliability, and drive faster, smarter operations as their catalog evolves.
2. Product Classification at Scale:
With large data volumes, classification is always a challenge. Manual classification of products was simply no match for the real amount of time invoice data flowing in. To solve this, we deployed our pre-trained auto-classification models that delivered with accuracy and efficiency.
The results were immediately tangible, and most impactful for upstream Enable processes, “Previously complex analyses that took days could now be completed almost instantaneously.”
Such is the impact of clean data that:
Conclusion:
We didn’t just help our client boost operational efficiency through automation — we also forged a powerful partnership with Enable, proving that when good data meets great strategy, the results can set new industry benchmarks.
