Fresh International develops Refresh™, enterprise software for cleansing, standardizing, classifying and enriching material, product and service data.
Our work in automated material data standardization began in 2006, when we filed our first patent in the field. Today, Refresh combines modern AI with two decades of proprietary Material Data Intelligence to help some of the world’s largest organizations improve the quality, consistency and usability of their enterprise data.
Headquartered in Switzerland, Fresh International supports multinational companies and public-sector organizations across North America, Europe, Asia-Pacific, South America and the Middle East.
What
Refresh is easy-to-use, preconfigured enterprise software that automatically cleanses, standardizes, classifies, enriches and deduplicates material, product and service data at enterprise scale.
It helps organizations create consistent, structured data across ERP, procurement, maintenance, design/PLM and other enterprise systems — quickly, accurately and with far less manual effort, while keeping core standardization controlled, repeatable and free from LLM-style hallucinations.
Why
Poor material data creates duplicate inventory, stock-outs, wasted working capital, fragmented spend, unsafe specifications and inefficient business processes.
Refresh transforms your data into a reliable corporate asset — helping you consolidate spend, reduce duplicate stock, improve search and reporting, support ERP transformation and strengthen compliance.
How
We combine modern AI, proprietary Material Data Intelligence and two decades of real-world automation experience.
The quality of AI results is ultimately constrained by the quality of the intelligence behind them. Two decades of clean, validated material data give Refresh the depth of intelligence to recognize what good results look like — and, just as importantly, what bad ones look like. Our technology has evolved continuously: from deterministic algorithms and highly vectorized text analysis through classical machine learning — including Bayesian methods, kNN, neural networks, support vector machines and decision trees — to the early adoption of GPT-based models and today’s LLMs.
Rather than replace proven methods whenever a new technology offers enhancement, Refresh combines the right technologies for each task. Modern AI is used where it adds capability, constrained by our strict ontology and validation framework, while deterministic and classical methods govern standardization where they provide greater accuracy, control and repeatability.
Our approach has always been software-first: automate what can be automated, continuously improve the underlying intelligence, make the easiest workflow the right workflow, and make sophisticated data standardization simple for the people who actually use it.
Who
Fresh International brings together experience spanning computational mathematics, classical machine learning, modern AI, enterprise software, procurement, engineering and ERP master data systems.
Our team understands both the technology and the business processes behind material data. Successful automation requires deep knowledge of how enterprise data is created, structured and used — and an understanding of which technology is appropriate for a particular problem, how to evaluate its output, and whether the answer is actually correct.
Combined with two decades of proprietary Material Data Intelligence, that experience gives Refresh an unusually deep foundation for automating and mastering enterprise master data.