Technische Unie is the largest technical wholesaler in the Netherlands. With 2,000 employees, Technische Unie unburdens its customers with specialist knowledge of products, total solutions and a wide product range. Technische Unie has 37 branches, 22 transfer points and 2 distribution centers in the Netherlands. Squadra Machine Learning Company did a proof-of-concept at Technische Unie to investigate whether errors in the data could be detected and possibly rectified using smart algorithms.
Challenge
At Technische Unie there are a number of challenges in relation to data quality. Namely, the data is, as with many companies, not always correct and not always complete. Squadra Machine Learning Company has investigated whether Technische Unie can be supported in improving data quality through a POC. In this POC, two SaaS products from Machine Learning Company were discussed, namely Powerimprove.ai and Powerenrich.ai.
Powerenrich.ai is a software that can automatically add extra features to product data by means of texts, images, and smart scraping of PDFs and websites. Powerimprove.ai is a software in which smart algorithms use a range of AI techniques to scan product data sets for inconsistencies, anomalies, duplications and improbable data values.
Solution
With the Powerimprove.ai software, Machine Learning Company has succeeded in checking whether the products of Technische Unie are classified correctly, whether contradictory data is present in the product data, and whether there are outliers present in the product data. With Powerenrich.ai we have succeeded in increasing the fill rate of the database of Technische Unie by automatically extracting product attributes from texts.
A Business Intelligence dashboard has been created to present the results of these SaaS products in a clear manner. On this PowerBI dashboard, Technische Unie can see its product groups, the fill rate per product group, and it is also possible to go into further detail and see the complete lists attributes per product. See an example below.
Technische Unie would like to be able to deal with the aforementioned problems in a smarter way in the future, which is why alerts have been set up. Those alerts may indicate incorrect data, immediately when new products arrive. Below is an example of a suspicious value, namely a very unlikely width, depth and length of 0 millimeters.
In addition, intelligent rules have also been added to the algorithm. In the example below you can see that there is 840 in the product name, where the 8 stands for the color and the 40 for the color temperature. These kinds of specific rules can be added and automatically converted to the correct product attributes.
Result
By using our Powerimprove.ai software, Technische Unie can now automatically correct datasets with incorrect data. This results in enormous time savings and benefits product data related processes. It also helps customers of Technische Unie because in this way the sales channels are provided with correct data.
The Powerimprove.ai software is offered as Software as a Service (SaaS), and is part of a group of SaaS products called the PowerSuite. For more information or an appointment, please contact us at info@powertext.ai or +31 (0) 85 065 3780.