In addition to conventional machine learning methods, we use Google Document AI and Google Gemini, cutting-edge deep learning models based on neural networks, to identify complex patterns even in large data sets. This llows us to present information and recommendations in a well-founded and dynamic way.
A major challenge in order processing is that orders reach us through various communication channels (e-mail, fax server, etc.) and in a variety of formats (PDF, text, Excel)—oft en in diff erent languages and with customer specific article numbers and descriptions. ChiruK@i automatically receives orders, translates them if necessary, and enters them accurately into our merchandise management system. Thanks to deep learning, ChiruK@i also recognizes new article patterns and reliably processes changes.
For trading companies, it is crucial to have suff icient goods in stock at all times while keeping capital commitment to a minimum. In our case, we also have to consider the diff erent delivery times within Europe. ChiruK@i now manages inventory planning independently: It compares current stock levels with orders and consumption data, detects trends, and takes into account delivery times and purchase prices from our European suppliers. From this, it creates cost-optimized order proposals. In addition to historical data, it also incorporates external factors such as seasonal fluctuations or delivery delays into its forecasts. Furthermore, ChiruK@i reviews all inventory figures daily and adjusts them in our merchandise management system to ensure short-term developments are addressed at an early stage.
We receive article data from customers and suppliers in various formats, languages, and often unstructured—spread across folders, e-mails, or desktops. Mapping identical or similar data and ensuring quick access for price and product comparisons involves a high manual effort. ChiruK@i is gradually building a central European article and price database. To do this, it reads all available information from suppliers and customers on a daily basis, identifies identical or similar articles, and intelligently consolidates them. This creates a comprehensive database that enables the automated processing of customer inquiries and the creation of tailored offers based on existing article data. Using demand forecasting, fluctuations in demand can be predicted much more accurately. The AI continuously adjusts these forecasts, reducing planning uncertainties.