Text Classification

The need:

Most communications and interactions with customers nowadays are done in Text format. There are multiple situations when finding the category a text belongs could increase the performance of a process drastically. At ILO they analyze thousands of news per year and categorize those to perform analytics on it. Know if a newspaper article is talking about labor initiatives, if these initiatives are an expansion of a plan are some of the more than 15 categories they are capturing.

This process takes considerable work for them, that is possible to automatize and scan much more news per year.

The solution:

  • Create several AI models based on the historical classification they have

  • Scrap the indicated web pages to find the news

  • Integration of the process with their current way to work

  • Create retrain pipelines to guarantee a continuous improvement of the models

Packaging boxes

The ROI:

  • Reduce the time a newspaper article takes to be analyzed

  • Increase the news analysis capacity on the team

  • Save time from the team to spend onmore interesting tasks