Denodo 9.4 boosts unified data access & AI at scale

Denodo has released Denodo Platform 9.4, adding features designed to help organisations move from AI pilots to production by improving governed access to enterprise data and boosting analytics performance.

The update expands unified access across structured and unstructured data and adds integration with vector databases. It also includes built-in support for the Model Context Protocol (MCP), which enables AI agents and clients to discover and query tools and data sources through a consistent interface.

Denodo frames the release as part of a broader shift in enterprise AI, as teams move beyond proof-of-concept work to focus on repeatable deployments, data quality controls, and operational governance. Data access remains a bottleneck for many AI projects, especially when information is spread across systems with different definitions and security models.

Unified data access

Platform 9.4 extends Denodo's ability to provide a single access layer across data types, including semi-structured and unstructured sources. The platform can connect to vector databases, which are commonly used to store embeddings generated from text, documents, and other content. These databases are often used in retrieval-augmented generation workflows that query enterprise content during AI interactions.

A shared access layer can reduce fragmentation in how AI systems connect to data and help enforce consistent definitions for business terms across sources. This is a common challenge for organisations that have accumulated separate data products and dashboards over time.

Built-in MCP support is positioned as another way to standardise governed access for AI systems. With MCP compatibility, an AI agent or client can discover and query live enterprise data through approved semantics and policies, rather than relying on bespoke integrations for each agent.

"As organizations move from AI pilots to production deployments, success increasingly depends on the intelligence and governance of the underlying data infrastructure," said Stewart Bond, Research Vice President, Data Intelligence and Integration Software, IDC. "Denodo Platform 9.4 reflects this shift by strengthening how enterprises unify structured and unstructured data access, embed governance into AI interactions, empower business users, and deliver consistent semantics across distributed environments. These capabilities are essential for building trusted, production-ready AI systems that can scale with operational demands."

Lakehouse performance

The release also introduces Lakehouse Accelerator, built on the open-source Velox execution engine. Denodo says it can deliver up to four times faster query performance, along with improved CPU and memory efficiency, for analytics and AI workloads.

Denodo describes Lakehouse Accelerator as an evolution of its embedded massively parallel processing approach, aimed at helping data engineering and platform teams handle more concurrent workloads without reworking architectures or moving data between systems.

Performance has become a central battleground as organisations try to serve both traditional reporting and AI-driven interactions from the same data foundation. Many enterprises now run mixed workloads-interactive analytics, batch processing, and AI retrieval-which can strain both the query layer and governance controls.

Business user interface

Platform 9.4 adds a conversational AI experience to the Denodo Data Marketplace, which provides a single point of access to enterprise data products. The interface supports multi-turn interactions, allowing users to ask follow-up questions and refine requests.

The interface is designed to show how questions are interpreted and answered and can prompt for additional input when context is unclear. Denodo positions it as a way for non-technical users to explore data without learning schemas or tools, while keeping governance controls in place.

Conversational access to governed enterprise data has become a common focus across data management vendors. Many organisations have tested generative AI internally but have struggled with permissions, auditability, and the risk of inaccurate answers when models lack reliable access to current information.

Production focus

Denodo is also highlighting agentic AI alongside generative AI. Agentic systems typically combine language models with tools that can take actions such as querying data, triggering workflows, or generating reports. This increases the importance of access control and policy enforcement, because agents that query live systems can expose sensitive information if governance is weak.

Denodo says its approach keeps governance teams in control and provides visibility into how agents and clients access data. It also emphasises consistent semantics to reduce confusion when similar fields carry different meanings across business units.

"Organisations are increasingly focused on turning AI ambition into real, operational outcomes," said Alberto Pan, Chief Technology Officer, Denodo. "Denodo Platform 9.4 is designed to support that shift by strengthening the data foundation across data teams, AI teams, and business users alike. By combining performance, governance, and intuitive access to live data, we help customers move from AI experimentation to trusted, production-ready AI that can truly differentiate the business."