Performancing Metrics

Estimated Reading Time: 4 minutes
Brainial Feature: automatic metadata extraction and DMS integration

Brainial Feature: automatic metadata extraction and DMS integration

The central storage of information and documents in a document management system (DMS) is becoming increasingly popular among companies. The amount of information within companies is growing every day. To organise and manage this information efficiently, many companies use a DMS. This, of course, also applies to tender documents. These documents must of course be stored in the right folder and with the right metadata (date, version, author, etc.) in the DMS, but this is often done manually. No Bid and Tender Manager is keen on having to do this for all tenders and so they spend a lot of time doing it or neglect to do it, especially if the annual volume of tenders or the number of documents per tender is high. The result is a waste of time or a poorly maintained DMS. Brainial has the solution!

Validated business case:

“A construction company measures a time saving of 75% per tender.

They process 100 tenders per year. On average, a tender consists of 150 documents.

This saves the company 45,000 euros per year.”

What is document metadata extraction? And which AI technologies are used?

Before we go to the solution, we first explain some of the concepts. Document metadata extraction is technology that ensures that information can be automatically extracted from files in one way or another and presented and processed in a structured data format. Before we can do this, we automatically transform all documents into an image. We then use Computer Vision to detect useful elements. Computer Vision is based on deep learning, and in Brainial’s case our models are specifically trained on documents and document formats. It is then converted into a structured data format or further processed with Natural Language Processing. A structured data format is a way of representing data that a computer can easily understand and that is consistent. Instead of large chunks of text, it is divided up and classified, a bit like a human using a marker to highlight a document. Brainial has developed its own models using Natural Language Processing to interpret the divided text, which allows us to detect and extract the metadata from the documents. 

Brainial Annotated Page

How does Brainial handle the metadata extraction of tender documents?

When a tender is received, all documents can be dragged and dropped into the Brainial solution at once. The analysis of the documents takes only a few minutes and all relevant information is extracted from the documents. If we look into our documents tab, we see an overview of all documents within one tender, together with the metadata and DMS folders predicted by the system. The metadata can be validated or corrected, and the system automatically keeps track of which documents have been processed, ensuring that nothing is overlooked and 100% accuracy can be achieved. When metadata is corrected, a sneak peak of the document can be viewed for validation at the push of a button. Documents can be selected and sent to the DMS where they are automatically and with the right metadata placed in the right folder. We have developed additional useful features such as filtering and searching for documents, bulk editing and bulk integration with the DMS. We also show the original folder structure in which the documents were delivered, and you can easily switch to the folder structure of your organisation. All in all, this is a set of functionalities that will make many Bid and Tender Managers happy. 

Brainial Document Tab

Benefits

It may not be our most obvious feature, but it is a feature that takes away a lot of stupid work, saves a lot of time (about 75% compared to the old situation), saves a lot of money (about 45K euro per year) and increases the quality and data integrity of the DMS system. But perhaps the most important benefit is that it leads to higher employee satisfaction. 

Curious about what other features we have? Take a look at our Product page. 

Want to see how it works? Schedule a demo and we will show you the benefits of using Brainial in your tender process.

Author: Taco Hiddink is the co-founder and chief technology officer at Brainial. He leads technical strategy and r&d team. Taco has over 10 years experience in data analytics and data engineering while working at a fast growing data analytics boutique company that was bought by private equity firm Waterland . Taco holds a Bachelor degree in Industrial Engineering & Management from the HU University of Applied Sciences in Utrecht.