Classification and automatic sorting of CEMs at the Italian National Institute for Social Security (INPS)

This information was sourced from the European Commission, Joint Research Centre (2026): PSTW: Public Sector Tech Watch latest dataset of selected cases. [Dataset] doi: 10.2905/JRC.J1MWC9R PID: http://data.europa.eu/89h/e8e7bddd-8510-4936-9fa6-7e1b399cbd92. Licensed under Creative Commons Attribution 4.0 (CC BY 4.0).

One of the most important channels by which citizens can communicate with INPS is via certified email (CE). On a practical level, when the INPS operator received the CE, he reads the text and any attachments, understands the subject matter (e.g., pensions, civil invalidity, redundancy payment), and then sorts it to the official in charge of answering the subject.

At the beginning of 2021, INPS started experimenting with the Open Source version of GPT-2, and then moved on to BERT, another Machine Learning model that also leverages the Transformers architecture and was developed by Google. The entire system was developed and executed in-house in the INPS Data Centres, precisely to fully satisfy GDPR and confidentiality requirements related to the sensitivity of citizens' data.

• It has accelerated the process of citizen communication with the Institution,
• It is active in 10 Italian cities of varying size and geographic area, including Rome, Milan and Naples and it has currently processed more than 1 million CEs. When fully operational, counting all territorial offices,
• It is estimated that between 30,000 and 40,000 working days per year will be saved.
• INPS was awarded by IRCAI, UNESCO's International Research Centre for Artificial Intelligence, in the world Top 10 projects supporting the 17 UN Sustainable Development Goals (SDGs).

Budget: Not disclosed