In business, people speed-read e-mails, reports, articles, publications, etc. and instantly understands what is relevant to themselves.

Try to think about the intellectual processing needed to do this kind evaluation; a person needs to understand the language of the text, and they need to understand the content and context of each sentence. They also need to understand the goals and priorities of the business and apply it to their evaluation. In addition, a person must be motivated, fed and filled with several liters of coffee every day.

Imagine using a computer to perform the same evaluation and let the machine decide what is important to you; How would you program it to do this task?

This was the challenge we gave to two of our interns from the American University in Bulgaria. Within 3 months they had to parse and analyze the given text correctly, extract facts and meanings and contextualize this in ways similar to how a human would do it. This week they presented the results of their 3-month-long project.


Martin Nestorov & Dimitar Kumanov in action, presenting their intern project on NLP at the Documaster HQ in Oslo in front of the Documaster Board and the Documaster team

What our two NLP pathfinders demonstrated was a small application they had developed. The application can analyze news stories from a database of about 25.000 articles and decide what the important elements are in the articles, associate a sentiment score to all the elements and highlight all keywords relevant to a specific context. In addition, the program can learn and self-improve with the help of user feedback and additional data, and then repeat iterations on the data when new ‘knowledge’ is acquired.

Of course, doing these kinds of student projects is not just for fun. It is an integral part of our commitment to recruit only the best and brightest developers. And equally important, to find new and innovative ways to apply NLP, ML and other aspects of Artificial Intelligence (AI) into the world of information archiving, because we believe traditional Records Management Systems have for too long been dependent on people making decisions and doing the right things. For decades, dedicated information workers, document controllers and managers have been policing and overseeing company processes to ensure compliance. In Documaster we believe that software can replace many of these tasks.

We strongly believe in the concept of letting unpolished, highly talented and creative students work inside Documaster R&D. We believe that fearless, young talents should be allowed to challenge what we do and thereby make us better, smarter and stronger. Documaster has already taken steps towards implementing AI in our products, but we are only at the very beginning. I predict that in less than a year from now, Documaster will be able to announce a new generation of AI-powered products that will set the new standard for Information and Records Management Systems. I’ll keep you posted.