Automatic Document Classification
ICT & Business
Client company:BDO
Wissam Alhamwi
Jurre Foekens
Wouter van de Burgt
Leandro Veldhuis
Project description
Main research question: How can documents automatically be classified on types using artificial intelligence for BDO while keeping law and regulations into account?
The goal of the project is to help automatically classify documents for BDO based on the document type. For example, if a document is added to a pipeline, from that moment onwards the document type, sensitivity, language and sentiment needs to be classified to the document.
Context
BDO is a company that helps every kind of company by looking at the business from a different perspective. The new economy changes faster than ever, and for that reason, it is necessary to stay innovative. This means that it is important to look with an open mind into opportunities. BDO does this with the goal of making their clients stronger, more agile, and more successful. When the trajectory is started, BDO has the role of advisor and independent accountancy.
BDO has the mission: to add security to organizations, entrepreneurs, and society by delivering top quality always and everywhere. BDO makes a difference through their data and technology-driven enterprising people from different backgrounds.
Results
We built a software solution in Python which is deployed in a so-called Azure Function. This Azure Function is again connected to a SharePoint environment. As soon as a new document is uploaded to SharePoint, our Azure Function is triggered by Power Automate, and the document classification starts.
About the project group
We are a group of ICT & Business students with multiple expertises. Some group members are specialized in e.g. cyber security and others in artificial intelligence.
The project we worked on has taken roughly 16 weeks. We have faced multiple challenges, but were able to overcome them all before the end of this semester
