Dell Technologies said on Wednesday it is expanding functions of the Dell AI Data Platform so companies can secure trusted, up-to-date context from their own data and provide it to applications and AI agents.

The company said much of the bottlenecks that arise as companies scale AI at large scale come not from the models themselves but from data. Core corporate information is scattered across different environments, including files, databases, cloud services and distributed systems across regions. Most data was created before the concept of AI agents was widely known, and is not classified or organized so that AI can read and understand it.

As a result, Dell said, AI agents must re-explore and combine multiple systems and data sources to produce answers rather than immediately using information that already exists inside a company. That process unnecessarily consumes many tokens and computing resources. As query processing steps increase, costs rise, response times slow and the reliability of results can also decline, it said.

The Dell AI Data Platform supports integrated operation of all layers from data to AI applications within a single platform, allowing customers to trace AI-generated results back to their existing governed data.

Its storage moves core data to an accelerated computing environment without bottlenecks, and data orchestration connects data and enables AI to use it. Through analytics and search functions, it transforms and indexes data. It also supports document parsing, embedding and reranking by using the NVIDIA Nemotron Retriever model.

AI agents repeatedly refer to the same documents, data histories and decision-making information, but in existing data environments it is often necessary to re-identify the meaning and relationships of data each time a new request comes in.

Dell also announced a Unified Semantic Layer, an Enterprise Knowledge Graph and Knowledge Agents, aiming to address these issues.

The unified semantic layer, knowledge graph and knowledge agents build context based on data, keep it up to date as business changes and share it across all applications and agents, without being dependent on a specific AI model, data or storage vendor.

Because it holds valuable and sensitive information, it is kept on a platform within a company's own data center.

Yoo Sang-mo (유상모), president of Dell Technologies Korea, said, "Being able to access data and being able to actually use data are different. Even if an AI agent finds certain information, if it cannot judge what that information means, how it is connected to other data, and whether it is trustworthy information, that agent is not intelligent but simply fast." He added, "The Dell AI Data Platform gives consistent and reliable context to enterprise data so AI agents can use it."