June 2026 Summaries

As software evolves, the role of the user is shifting from humans to autonomous agents like Claude and Codex, which can perform tasks traditionally done by people, such as logging into tools, updating data models, and generating insights at a much faster pace. This transformation is evident in business intelligence (BI), where agents can now manage entire workflows—data modeling, exploration, and presentation—traditionally requiring a team of data engineers and analysts. Despite this shift, the existing infrastructure supporting human users, including data model SDLC, security, governance, and semantic layers, remains crucial for ensuring agents operate effectively and reliably. Cube is adapting to this change by enhancing programmatic access to its platform, allowing agents to interact with its semantic layer and perform tasks seamlessly, indicating a future where agents will increasingly handle BI operations while maintaining data integrity and trustworthiness.

Cube Evals is a newly launched feature that addresses the challenges data teams face when AI agents answer business questions using a company's data model. As AI agents become integral to production systems, ensuring the accuracy of their responses is critical. Cube Evals allows teams to create evaluation cases—pairing natural-language questions with known-correct answers—and run these against the AI agent to obtain an objective accuracy score. This process helps identify discrepancies between the agent's output and the ground-truth data, allowing for targeted improvements. The evaluation cases are stored in the data model repository, ensuring that testing is integrated into the existing workflow and can be easily managed alongside code changes. This approach provides a deterministic grading system, ensuring consistent and reproducible results, and can be enhanced with optional model-based grading for more nuanced assessments. By automating what was previously a manual and error-prone process, Cube Evals streamlines validation in AI Studio, making it a native part of the development and deployment workflow for organizations using Cube.

MCP Connectors have been introduced to enhance the Cube agent's ability to provide context-rich answers by integrating with existing tools like Notion, Linear, Sentry, and CRMs, allowing it to access information beyond raw data metrics. These connectors operate on the Model Context Protocol (MCP), an open standard that enables the agent to use external tools in conjunction with semantic-layer queries, without altering the foundational data definitions and access policies. By leveraging outbound integrations, the Cube agent can draw upon information from various sources to deliver comprehensive insights into business questions, such as understanding the reasons behind changes in key metrics. Administrators can manage these connectors through a directory of vetted integrations or by setting up custom connections. This system ensures that the agent respects existing permissions and access controls, while allowing users to query and build semantic models with enriched metadata from external sources. Overall, MCP Connectors expand the agent's capabilities by bridging the gap between numerical data and qualitative context, facilitating more informed decision-making and seamless integration of analytical workflows into everyday business processes.

Embedded Agentic Analytics is a new offering from Cube, designed to integrate AI-driven analytics seamlessly into products using the Cube semantic layer. Unlike traditional embedded analytics, which primarily involve dashboards and chart components, this new solution caters to modern customer expectations of interacting with data through natural language queries and obtaining precise answers without leaving the product. It addresses the limitations of past analytics stacks by ensuring AI agents have the requisite context to produce accurate results, thanks to the underlying semantic layer that contains definitions, measures, and policies. Cube's approach maintains a balance between governance and flexibility, allowing for ad-hoc calculations within a governed model. The offering includes four embedding options—Analytics Chat API, Chat and Dashboard iframes, Creator Mode, and Core Data APIs—all running on the same semantic layer, enabling easy transitions between options without rebuilding the data model. A notable example is Brex, which successfully implemented an AI-powered financial reporting workspace using Cube, demonstrating the practical application and benefits of this approach in handling multi-tenant architectures and scale. Cube invites existing and new customers to explore these embedding options, promising further enhancements and integrations in the future.

Cube has introduced Cube Agent Skills as a tool to streamline repetitive analytical workflows by capturing them as reusable skills within the Cube platform. These skills are essentially markdown files integrated into the Cube project, containing metadata and instructions that automate multi-step analytical processes. This innovation allows recurring tasks, such as weekly reports or customer churn analyses, to be executed consistently without manual re-entry. The skills are accessible via buttons, a slash menu, or automatic matches to user requests, and they inherit existing governance structures like code review and version history. While the initial release focuses on core functionality, future updates are anticipated to include features such as scheduled tasks and extended external actions. Cube Agent Skills are currently available, offering a practical solution for teams to transform routine manual processes into efficient, shareable artifacts.