Data is everywhere, and making sense of it is more critical than ever. Yet, organizations continue to face challenges with increasing data complexity, governance hurdles, and the difficulty of providing unified access to insights. While traditional Business Intelligence (BI) tools have benefited from semantic layers, bridging the gap between complex data structures and business insights often still requires significant technical expertise.

The rapid advancements in Artificial Intelligence (AI), particularly Large Language Models (LLMs), offer a transformative opportunity for data analysis. However, for AI agents to deliver reliable and trustworthy insights, they need to accurately interpret and interact with corporate data. This is precisely the problem Cube Cloud's new D3 agentic analytics platform is designed to solve. D3 leverages the power of Cube's robust semantic layer to empower both data consumers and data stewards with augmented self-serve capabilities.

The Essential Role of the Semantic Core

At the heart of D3 is Cube's semantic layer. For AI agents to be effective, they need more than just raw data or database schemas. They require a foundational understanding of the business context, terminology, and the intricate relationships within the data. The semantic layer serves as this crucial structured bridge between complex data sources and business requirements. It transforms the technical language of data into the understandable vocabulary of the business, much like a Rosetta Stone translates ancient text.

The semantic layer provides a single source of truth for business metrics and dimensions, ensuring consistent definitions and reducing ambiguity. By embedding metadata and standardizing business logic, it offers the deep contextual understanding necessary for AI agents. This layer defines concepts, relationships, and rules that guide agents in interpreting user queries and generating accurate results. Without this governed, business-friendly layer, AI agents querying raw data directly are prone to errors, miscalculations, and "hallucinations".

Cube's semantic layer includes components like metadata repositories, schema mapping, taxonomy/ontology management, and robust query engines. It embodies the principles of a knowledge graph, allowing D3's agents to grasp the business relevance of data. Critically, the semantic layer includes a compiler that translates high-level business requests into correct and optimized SQL queries, offloading this complex task from the AI agent and drastically reducing error risk. This approach is fundamentally superior to solely relying on text-to-SQL methods for enterprise data.

How D3 Works: Agentic Analytics Powered by Semantics

D3 is designed as an agentic analytics platform part of Cube Cloud. Its core idea is to allow users, regardless of their SQL knowledge, to interact with data using natural language. The platform augments users with analytics agents grounded in Cube's semantic layer to eliminate technical bottlenecks, enabling rapid, independent data exploration and driving insight velocity.

Key Components: Agents, Spaces, and Rules

D3 introduces several key components to enable this agentic analytics experience:

  • Analytics Agents: These are the AI-powered entities that assist users. They search the semantic model and can learn from experience and data model context. They support the selection of various LLMs, including the ability to Bring Your Own Model (BYOM). Agents can be restricted to specific data views for enhanced security and focus.
  • AI Data Analyst: This agent is specifically designed to assist Data Consumers. It creates Semantic SQL queries, visualizations, and data apps. Results are provided in the chat interface as Semantic SQL Tool Call Messages (SSTCMs). SSTCMs are immutable but can be opened in the Data Explorer for further investigation.
  • AI Data Engineering Agent: This agent supports Data Stewards. It helps in creating and optimizing semantic models, such as building a model from existing data marts or adding measures and dimensions.
  • Agent Spaces: These are logical groupings used to manage agents and related resources. Spaces support a role-based permission model for simplified management and facilitate the creation, viewing, and deletion of agent groups.
  • Rules: Defined guidelines within a space manage agent behavior and data interaction. This includes "always rules" and "agent requested rules". Establishing rules is crucial for guiding the AI on how to query specific data points, like snapshot measures.
  • Certified Queries: These are SQL queries that have been reviewed and approved by admins, which agents can use within a specific space.

Other key components include the Analytics Chat Interface (for natural language interaction), the Data Explorer (for interacting with SSTCMs and Reports), Reports (saved analytical outputs derived from SSTCMs), and the Data Assets Pane (listing available semantic views, reports, and data source tables).

Using and Setting Up D3

The D3 experience is designed around an interactive chat interface and the Data Explorer.

For Data Consumers: Users can start by typing analysis requests in plain language into the chat interface. The AI Data Analyst generates queries and visualizations, presenting results as SSTCMs in the chat. Users can then open SSTCMs or Reports in the Data Explorer to view the underlying query, results, and visualizations. The Data Explorer allows users to modify the SQL query, run it, and send the updated query back to the chat to refine their analysis or provide feedback to the agent. Valuable results can be saved as Reports, which can include names, descriptions, and context, and can be shared. The Data Assets Pane provides navigation to available assets.

For Data Stewards: The AI Data Engineering Agent assists in creating and optimizing semantic models, potentially generating YAML code for cubes and views from existing tables.

For Admins (Setup and Management): Access the Admin Panel via the Main Menu sidebar. Here, you can manage Users (viewing, changing roles), Global Settings (including the BYOM section to configure custom AI models by specifying name, provider, model, and potentially access keys - note that MCP server settings are coming soon), Agent Spaces (creating, viewing, and deleting spaces), manage Certified Queries within spaces (viewing, deleting), and manage Rules within spaces (adding rules with type, prompt, and description). Admins also manage individual Agents(creating agents by name, model, and space; deleting agents; clearing chat history) and can restrict agent visibility to specific views in the agent settings.

For initial customers, the Cube team will provide guided setup, collecting views or tables and sample questions, enabling D3, and scheduling a demo. A recommended onboarding for general availability involves the Platform Admin using the AI Data Engineer to create the initial model and creating verified reports for context and examples.

Availability

D3 is an agentic analytics platform that is part of Cube Cloud. It is being rolled out, with the initial focus on supporting existing and new Cube Cloud customers via a guided process.

Release Features & What's Next

The initial release introduces core D3 capabilities, including features available in public preview:

  • Core platform functionality: settings, users, agents and spaces management.
  • AI Data Analyst capabilities.
  • Analytics Chat interface.
  • Data Explorer & Reports Management.
  • Semantic Model Development assistance via the AI Data Engineering Agent.
  • Data Assets Navigation panel.

But this is just the beginning. We are actively developing exciting new features, including those planned for future releases:

  • Reports sharing capabilities.
  • Interactive data apps.
  • Pro-active semantic model development suggestions from agents.
  • Structured Queries interface within the Data Explorer.
  • Future support for AI Data Scientist capabilities.

Conclusion

The convergence of AI and BI is reshaping how organizations interact with data. Cube D3 stands at the forefront of this shift by combining agentic AI with the power of a robust semantic layer. By providing a governed, contextualized bridge between complex data and business users, D3 empowers organizations to achieve greater data accessibility, accelerate insight generation, and ensure the trustworthiness of AI-driven analysis. This approach positions D3 as a key component for unlocking the full potential of AI in data analytics workflows.

Contact sales to learn more.