Banking sector

Automation of document processing

The challenge

Our client receives a huge load of (unstructured) financial documents on a daily basis. These documents have to be dispatched to the right team and relevant data points have to be extracted. This is a significant manual effort that is prone to human error. This workload will only increase towards the future due to compliance regulations.

Our approach

Working closely with the client, Dataroots developed a topic/text classification model that is able to automatically dispatch the documents to the right team for processing. After dispatchment these documents are processed by a text analytics (NLP) engine to automatically extract relevant data points from these unstructured documents.

The outcome

By integrating our solution, the client is now able to effectively automate a major portion of his manual workload. This has allowed for a large scale-up in their document processing efforts at a fraction of the cost while significantly decreasing the element of human error.