Private Chat AI

Project Overview

#

The Private Chatbot AI project was designed with the goal of developing a private approach to the use of Large Language Models (LLM), integrating them with corporate data in a protected environment, without transferring such information online or sharing it with servers external to the company, especially if controlled by non-EU entities. This approach is fully aligned with the principles of the GDPR regulation and the requirements of the AI Act.

Project Results

#

The goal was fully achieved: during the project, a modular, flexible, and secure system was developed, designed to meet the needs of businesses and contribute to the goals of the smart factory and sustainable development. The result lays the foundation for advanced technological evolution, particularly in the context of Made in Italy. The system is modular and consists of several functional blocks: it required constant research activity, also in light of the rapid developments in the field of LLM and the growing awareness, on the part of companies, of the importance of adopting private and controlled solutions. Its modularity allowed the development of concurrent functionalities and the capture of innovations as they arose. Thanks to what has been developed, it is now possible to interact via a web chat with heterogeneous corporate data (documents, databases, text files), using different language models hosted locally or on private-controlled European clouds.

Technological Impact

#

For SMEs

#

• Full control: Data always under corporate control
• Customization: Specific adaptation to business processes
• Scalability: Modular growth based on needs

For the manufacturing sector

#

• IoT integration: Direct connection with industrial sensors and machinery
• Supply chain management: Automatic optimization of the supply chain
• Predictive maintenance: Preventive analysis of failures through AI

Future Prospects

#

PrivateChatAI represents the basis for further developments in the field of private and secure AI. The results of the project are already fueling new research and developments for:

• Extension to new industrial sectors
• Integration with existing ERP and CRM systems
• Development of multimodal capabilities (voice, images, documents)

October 2025: first commercial products

#

The PrivateChatAI project has already generated its first commercial product: ArisQL, an enterprise solution for integrating natural language to SQL conversion in corporate products.

ArisQL represents the concretization of the research conducted during the project, transforming the developed technologies into a market-ready product, designed to ensure accuracy, security, and privacy.

November 2025: the project among the best in the FVG Region

#

At our headquarters at BIC Incubatori FVG, we were visited by the representative of the FESR Projects Commission Joanna Olechnowicz, Dr. Marina Valenta, and architect Lino Vasinis of the Central Finance Directorate of the Autonomous Region of Friuli Venezia Giulia to learn about our Private Chat AI project, highlighted as one of the best in the region!

December 2025: new project funded

#

Starting December 1, 2025, and lasting 12 months, the project “AI for pre-operative classification support” will begin: built on the foundations of the Private Chat AI project, the project aims to evolve a patient classifier according to the guidelines of the American Society of Anesthesiologists.