Dell adds AI data tools for governed enterprise use

Dell Technologies has expanded its Dell AI Data Platform with new tools designed to give AI agents access to governed enterprise data.

The update centres on three additions: a Unified Semantic Layer, an Enterprise Knowledge Graph and Knowledge Agents. Together, they are intended to help organisations connect data across files, databases, cloud services and internal systems, then present it to AI applications with shared definitions and access controls.

Businesses deploying AI systems have faced a recurring problem when data is spread across older and newer repositories that were not structured for machine use. Dell's latest update aims to address that by attaching common meaning to data and mapping relationships across records, documents and other information sources.

The Unified Semantic Layer is designed to give structured and unstructured information a common business meaning. It can apply rules, definitions and glossary terms so different labels used in separate systems are recognised as referring to the same concept. It also allows users to import existing ontologies and classification taxonomies.

The Enterprise Knowledge Graph is intended to show how structured and unstructured data relate to each other. Dell said it uses metadata, lineage and query history to refine those relationships over time, and can bring together related tables, data products, multimodal data and vector indexes that a user or agent is permitted to access.

Knowledge Agents sit on top of that layer. Each agent is tied to a defined section of the knowledge graph and can be governed through rules covering guidance, data access, quality thresholds and spending limits.

NVIDIA technology features across the platform. Dell said NVIDIA Nemotron Retriever models are used for document parsing, embedding and reranking, while NVIDIA cuVS handles vector indexing and search. It is also enabling NVIDIA Auto-Ontology, which Dell described as an open-source library for building knowledge graphs from enterprise data.

Dell said the semantic layer, knowledge graph and knowledge agents remain inside a customer's data centre because they hold sensitive information. The approach is intended to help organisations keep context current and share it across applications and agents while avoiding dependence on a single model, data or storage provider.

Data path

Dell also announced changes to data processing and storage, with a focus on moving information more quickly between storage and compute systems. In internal testing, Dell said its Data Processing Engine, using NVIDIA cuDF on NVIDIA RTX PRO 4500 Blackwell Server Edition GPUs with Apache Arrow, processed data nearly four times faster on average than CPU-only systems across a range of workloads.

The same tests showed peak speed gains of up to 20 times on batch data-mining workloads, according to Dell. The company added that Apache Arrow allows jobs to query data in place, which can reduce preparation time between raw information and AI use.

On storage, PowerScale now supports up to 500 tenants in a single cluster, Dell said. The company also added mTLS over NFS to encrypt and authenticate file traffic, alongside more granular role-based access controls for managing each tenant.

Another addition is the Dell Storage Performance Tool, which is intended to help customers size AI infrastructure and compare S3-compatible object storage performance across training, inference and checkpointing workloads.

Services push

Dell is also expanding its implementation services for the platform's data and storage engines. These services are intended to help customers move from initial deployment into production use, including work on analytics, processing, search and orchestration.

Arthur Lewis, President, Infrastructure Solutions Group, Dell Technologies, said the changes reflect a broader shift in enterprise AI, where access to information is not enough if systems cannot interpret and trust what they retrieve.

"Data without context is just noise. Most enterprises have spent years making their data accessible. That's not the same as making it usable. An agent that can find a customer record but doesn't know what it means, how it connects to everything else, or whether it can be trusted isn't intelligent. It's just fast. The companies that solve it won't just deploy AI faster or run more agents. They'll get more out of the data they already have," Lewis said.

Jason Hardy, Vice President of Storage Technology at NVIDIA, said: "AI agents are only as effective as the data they can access, understand and trust, and NVIDIA accelerated computing and AI software help turn governed data into AI-ready context. By bringing NVIDIA cuDF acceleration directly into the Dell Data Processing Engine, Dell helps shorten the path from stored data to GPU-accelerated, AI-ready data that agents can use to drive real-world innovation."