November 20, 2020

Using AI to classify legal documents

A common practice in the Greek legislation system is the incorporation of various articles that modify existing legislation in new laws voted by the parliament. These articles can be irrelevant to the main content of the law to be voted on.

Due to this pattern, interested parties find it really challenging to keep up with the evolution of the effective law that affects their interest. Not to mention that failing to follow the current laws may result in really harmful situations for any organization.

To handle this difficulty Nomiki Bibliothiki, a leader in the field of law and economics, followed methods of coding laws and manually classifying new legal documents. A significant number of lawyers was required and for a two-phase process:

a) primary classification and
b) random control and verification.

Bear in mind that new legislations are being produced almost every day by state officials. Thus, there were considerable disadvantages in managing this situation without any Artificial Intelligence (AI) tools. The process required too many resources with increased costs. It was really slow, mostly ineffective, and there was an increased risk of errors.


The solution
The solution

How ahedd Helped in This Case – The Solution

The role of ahedd

ahedd is a Digital Innovation Hub for the economy of data and devices. The first thing to do is to identify the need of the client and conduct a business analysis. Then we should match this need with the technological competences of its partner, SciFY.

Working as a united entity, the client was provided with a single point-of-contact to build trust and facilitate all communication.

The solution

SciFY, in close collaboration with ahedd, investigated the whole process and involved managers and end-users in the analysis to determine that the solution would be effective and really used.The team concluded that SciFY had to create a “robotic” assistant that would search into thousands of texts, decide what they were referring to, and automatically classify them. It would also suggest “tags” that would assist in searching and make correlations with ease.

SciFY developed an Artificial Intelligence system and more specifically machine learning algorithms. The algorithms were programmed to extract the subject and relevant labels from the documents to classify them and add metadata for easier search.

The above were incorporated into the existing services of the client and the results exceeded the client’s expectations.

Case study recognition

Boston Consulting Group (BCG) featured this case study in “Harnessing the Power of AI in Greece“, published in September 2020. You may access the study in .pdf format here:

Other areas of application

A similar approach can be applied for bringing AI-empowered solutions to other professional services such as media, press indexing, customer profiling (banking, credit cards, public services), and data management.

What people said

quote icon

With this investment, we are reaping significant benefits for our company and our customers: we are saving precious time that the staff was up to now spending for the thematic classification of legislation and the enrichment with metadata. Our productivity has significantly increased. So, we can now focus more on the most value-adding parts of the job, where the human contribution is important. Also, the position of our company as a provider of legal information services is strengthened.

Nomiki Bibliothiki
star icon

Lets meet,
Lets meet,


contact icon


What's your next project? Let's build yours, together. Whether you want to talk about the next big thing or simply say hello, we'd love to hear from you.

Scroll to top