Contents
Grupo Gigas, the Spanish multinational specializing in cloud, cybersecurity, and communications services for businesses, announced the launch of Biblion, a generative AI platform designed specifically for private corporate environments.
The solution comes at a time of rapid adoption of generative AI tools, when organizations want to leverage the capabilities of large language models without compromising the confidentiality of their internal information or losing control over their strategic knowledge.
Subscribe to Mundo Ejecutivo to read exclusive content and receive benefits
The Use of AI in Businesses
In Mexico, 27% of organizations report having a clearly defined artificial intelligence strategy that is aligned with their corporate objectives.
However, only 41% have a general understanding of AI, without having yet developed a business case.
In addition, another 31% have expressed interest in implementing this technology, although they are still unsure about how to go about it.
According to recent industry studies, more than a third of the information that employees enter into public AI tools is sensitive in nature.
Information Security
This solution enables companies to search, summarize, and interact with their own internal documentation using natural language, leveraging advanced language models (LLMs) combined with Retrieval-Augmented Generation (RAG) techniques.
Unlike general-purpose public models, the platform generates responses exclusively based on the organization’s authorized document corpus, avoiding the use of external sources and reducing the risk of inaccurate or unverified responses.
Download the February issue of Mundo Ejecutivo magazine, featuring Gonzalo Araújo on the cover
Generative AI with full data traceability and control
One of the key features of Gigas’ generative AI is its ability to provide traceable responses. Each generated response includes direct references to the specific documents and web pages used in its creation, allowing users to verify the original source and providing a level of reliability that is particularly valuable in legal, technical, or regulatory departments.
The solution is designed to transform large volumes of documentation—such as technical manuals, guidelines, internal procedures, knowledge bases, and regulations—into accessible information in seconds.
Similarly, the platform supports PDF files, includes version control and role-based permissions, as well as a search history and usage statistics.
It also supports multilingual environments, allowing you to upload documentation in different languages and generate responses in the language in which the query is submitted.
RAG Architecture in a Private Cloud
Biblion is built on an architecture that integrates vector databases, semantic vectorization techniques, and intelligent chunking, optimizing the retrieval of relevant information. This approach allows for the updating of document knowledge without the need to retrain the entire model, ensuring agility and accuracy.
The solution is deployed as Software as a Service (SaaS) on Gigas’ private cloud, in a dedicated environment for each customer. Customer documentation remains entirely within Gigas’ infrastructure. Queries and responses are logged in the customer’s private environment, and the information sent to the language model is not used to train third-party models.
The system includes two-factor authentication (2FA) and encryption in transit and at rest, and can be integrated into closed corporate environments, such as intranets or VPNs, when required by the organization.
According to Víctor Guerrero, CEO of Grupo Gigas, companies need to integrate generative artificial intelligence into their processes without compromising the sovereignty of their data.
“Biblion combines the power of language models with a private, traceable, and controlled environment, aligned with the security requirements of corporate environments.”
This generative AI marks a turning point in the evolution of the private cloud toward sovereign generative AI models, in which organizations can integrate large language models (LLMs) into their critical processes without relinquishing control, traceability, and governance over their corporate knowledge, thereby establishing data sovereignty as a strategic pillar of digital transformation.
Follow us on Google News to stay informed
