A company specializing in the design and construction of cranes for the maritime industry needed to make its extensive technical knowledge easily accessible. This information was spread across multiple archives and business systems, making it essential to simplify information retrieval and support the daily operations of different departments.

introduction

Trasformare il patrimonio documentale in conoscenza accessibile

In the marine crane design and manufacturing industry, every project involves managing large volumes of information: technical specifications, drawings, structural calculations, certifications, revisions, and communications with customers and suppliers.

Over time, this content becomes a valuable corporate knowledge asset, bringing together expertise, experience, and solutions developed throughout different projects. However, when information is scattered across multiple archives and systems, retrieving a document or identifying a previously implemented solution can become a complex and time-consuming task.

AI-powered Knowledge Management helps overcome these challenges by making information more accessible and facilitating its reuse across business processes.

Trasformare il patrimonio documentale in conoscenza accessibile
The Context

Un archivio tecnico complesso, distribuito tra diversi sistemi

The client is an industrial group specializing in the design and manufacturing of cranes for marine applications, characterized by complex projects and a high degree of customization.

Documentation was distributed across file systems and SharePoint, organized into different hierarchical structures based on departments, projects, and revisions. This made information retrieval challenging, particularly when data from previous projects needed to be accessed quickly to prepare new proposals or respond to customers' technical inquiries.

The objective was therefore to centralize access to corporate knowledge while preserving the structure of existing archives and ensuring the confidentiality of information.

Un archivio tecnico complesso, distribuito tra diversi sistemi

The Challenges

Integration of heterogeneous information sources

Overcome the fragmentation of information distributed across file systems and SharePoint by creating a unified access point to corporate knowledge.

Advanced search and content understanding

Enable users to find specific information by asking questions in natural language, receiving relevant and verifiable answers without relying solely on traditional keyword-based searches.

Navigation of the document structure

Enable the system to understand not only document content but also its hierarchical organization, identifying file locations and relationships between projects, folders, and documents.

Security and access control

Ensure that each user can access only the information authorized for their role, protecting the confidentiality of technical and commercial data.

The Solution

KI Engine: an AI platform to unlock the value of corporate knowledge

To address these needs, artea.com developed a private Knowledge Management platform based on KI Engine, designed to transform corporate documents and information into a knowledge base accessible through Artificial Intelligence.

The solution integrates content from SharePoint and file systems, automating its acquisition, updating, and semantic organization. Thanks to Generative AI and Retrieval-Augmented Generation (RAG) technologies, users can query the document archive in natural language, receiving relevant and verifiable answers with references to the original sources.

The platform simplifies the retrieval of technical information and documentation from previous projects, while also supporting the preparation of commercial proposals.

Specialized AI agents were also configured for six business departments, with role-based access managed through Active Directory and Single Sign-On (SSO), on a Microsoft Azure architecture designed according to privacy-by-design principles.

One of the project’s distinctive features is KI Engine’s ability to interpret not only document content but also the structure of the corporate archive.

Unlike traditional RAG systems, which retrieve information from indexed content, the solution developed by artea.com enables the language model to navigate the document hierarchy, identifying projects, folders, and archives, locating files, and determining the number of documents available.

KI Engine thus goes beyond the limitations of basic semantic search, making corporate knowledge organized, contextualized, and navigable to support operational and decision-making processes.

The implementation of KI Engine has made it possible to centralize and provide access to approximately 900,000 pages of technical documentation, simplifying information retrieval and access to previously developed engineering knowledge.

The platform, used by 45 users across 6 business departments, now supports tender preparation and the management of customers' technical inquiries through semantic search and document navigation capabilities.