Document Processing · Enterprise Knowledge Search
BNP Paribas needed a secure, standardized way for its entities to access and integrate large language models so it could industrialize generative AI use cases across the Group while meeting privacy, security, and cost-control requirements.
Reported impact
The platform gives BNP Paribas a common, secure foundation for scaling generative AI, accelerating deployment of new projects, improving integration of AI into existing processes, and applying shared technical and security standards across entities.
BNP Paribas deployed an internal "LLM as a Service" platform to give business units and functions unified access to large language models through a standardized, secure interface. The platform is operated by the Group's IT teams and hosted in BNP Paribas data centers, where specialized GPU infrastructure provides the processing power to run the models inside the bank's own environment. Through this layer, entities can select open-source models or models from Mistral AI that fit their needs and integrate them directly into existing tools and processes. The shared platform serves current use cases such as internal assistants, document generation, and document information retrieval, while enforcing common technical and security standards. BNP Paribas says this common foundation accelerates deployment, supports progressive roll-out across entities, and helps share and limit costs while aligning generative AI usage with business and compliance requirements.
Group entities access the internal LLM as a Service platform through a standardized secure interface; the platform routes requests to selected open-source or Mistral AI models running on GPU-equipped infrastructure in BNP Paribas data centers, and outputs are integrated back into internal tools and business processes for assistants, document generation, and document retrieval use cases.
"BNP Paribas has now deployed an internal LLM as a Service platform, designed to provide the Group's entities with unified access to large-scale language models (Large Language Model - LLM) in a secure environment within the Group's own infrastructure."
"Several generative AI use cases are already in production or experimentation within the Group's businesses, such as internal assistants, document generation, and information retrieval in documents."
"The implementation of the platform allows for scaling up by facilitating the deployment of these use cases within a common framework and accelerating the implementation of new projects."
"The platform is based on a shared infrastructure hosted in the Group's data centers, equipped with specialized computing capabilities (GPU). It provides access to a selection of open-source models, models from Mistral AI, a partner of the Group"
"These models can be directly integrated into tools or processes."
"This shared infrastructure, operated by the Group's IT teams, is set to accelerate the development of generative artificial intelligence (AI) use cases to better serve employees and clients."
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Mistral Sample Apr 24, 2026 Ardian Ardian needed to analyze large volumes of confidential due diligence material faster, but its teams were overwhelmed by scattered internal documents, NDA constraints, and traditional machine-learning approaches that did… 18,000+ documents since inception GAIA · Mistral models 
MongoDB Sample Oct 8, 2026 Navisio AI Rapid adoption of Navisio AI's platform triggered severe latency spikes, elevated CPU and RAM consumption, and costly vertical scaling. As it prepares to scale to 1,000 by the end of the year, and eventually toward 5,000 users, the team is… MongoDB Atlas · MongoDB Atlas Vector Search 
Google Cloud Sample Archived · Sep 27, 2026 SIGNAL IDUNA SIGNAL IDUNA needed to spread AI safely across a large insurance workforce, make a century of internal policy knowledge searchable, and let teams build their own agents without creating a fragmented, insecure shadow-AI… 80% weekly active adoption rate with a +45 NPS in first five months Document AI · Gemini · Gemini Enterprise Agent Platform 