The analytics landscape is undergoing a rapid evolution. While dashboards still aren’t dead, analysts are increasingly turning to chat interfaces to understand and explore their data. Even as organizations are figuring out how to make these chat interfaces more trustworthy by building an AI-ready semantic layer, a new frontier for analytics has emerged: the rise of analytics agents to investigate data, make recommendations, and trigger workflows.
In this blog, we’ll discuss key updates from the 2026 Databricks Data + AI Summit and how they’re enabling this new analytics paradigm.
Current State: From Dashboards to Conversational Exploration
No longer relegated only to proofs of concept (POCs), AI is already reshaping how users interact with data. Across organizations, we’re seeing two clear patterns emerge.
- AI for exploration: Conversational interfaces that allow users to ask follow-up questions, drill into metrics, and iterate without writing SQL
- AI for interpretation: Automated summaries, anomaly explanations, and narrative insights layered on top of dashboards
Databricks had already been investing heavily in this space with offerings like Genie Spaces and Metric Views, which aimed to make conversational analytics more reliable by grounding responses in governed business logic.
However, while chat-based analytics improves accessibility and reduces the time to insight, it cannot take autonomous action, making it fundamentally reactive. This raises an important question: if dashboards and chat interfaces are helpful but not sufficient, what should the future of analytics interaction look like?
The Rise of Agentic Analytics
Agents extend the analytics toolkit meaningfully. They can take action proactively, which unlocks the ability to investigate issues, assess findings, and address the issues autonomously. Rather than replacing existing approaches, agents become a key addition to a broader analytics ecosystem, working alongside other interaction modes. Analysts can then choose the mode that works best for their use case:
- Dashboards for standardized, repeatable monitoring of KPIs and operational metrics (without forcing your users to ask the same question every time)
- Chat interfaces for ad hoc exploratory analysis, iterative questioning, and rapid insight generation
- Deterministic workflows for recurring, standardized processes like reporting, alerts, or data transformations
- Analytics agents for recurring processes that require inference (e.g., choose a topic to investigate, reason on findings, and act on behalf of users)
Importantly, these modes should be fluid. An analyst might start with a dashboard, use a chat interface to ask follow up questions, and then delegate deeper investigation to an agent. As part of its output, the agent might create a visualization that can be saved back to the dashboard for future use.
Moving from isolated interfaces to an agent-centered, multi-modal experience sets the foundation for the next wave of analytics innovation.
2026 DAIS Announcements for Agentic Analytics
The 2026 Databricks summit revealed that Databricks is taking a holistic approach to enabling agentic analytics. Rather than treating agents as a standalone feature, Databricks is investing across multiple layers of the stack.
Access Layer
Databricks is crafting user-oriented experiences to interact with agents. Genie One (formerly Databricks One) introduces a more unified conversational interface, designed to serve as the entry point for analytics across personas. Integrations into tools like Slack and other enterprise applications allow users to interact with data without switching contexts. Genie One makes it easy for users to create and use skills and agents, schedule tasks, and create documents. As Genie One can leverage domain-specific Genie Agents, cross-domain questions become easier and more reliable.
Agent Layer
Genie Spaces are getting an upgrade—and a rename to match. Now called Genie Agents, they support write capabilities (not just read), MCP reads and writes, sub-agents, and skills. Teams are able to package, share, and reuse domain-specific analytical agents. This positions Genie Agents as more than just chat interfaces.
Semantic Layer
A major focus at DAIS 2026 was the semantic layer. As AI-driven analytics becomes more prevalent, ensuring agentic answers are accurate and context-aware takes a front seat.
Databricks is taking a multi-faceted approach to upgrading its semantic layer:
- Genie Ontology is a permission-aware, page-ranking knowledge graph. It integrates both the governed semantic layer (explicitly defined domains, metric views, and glossary pages) with business knowledge from external sources (e.g., SharePoint, Google Drive, Confluence, etc.). In doing so, it aims to address the staleness and drift that can make a governed semantic layer degrade over time.
- Domains are business-aligned categories. Metric views, glossary pages, and other resources can be tagged to specific business domains, making it easier for Genie One (via Genie Ontology) to surface relevant content for a given chat or task.
- Glossary lets you define concepts, terms, and taxonomies that help agents and people understand your business. Glossary pages can link to each other and to assets, thereby capturing relationships.
- Metric Views ensure consistent definitions of KPIs, supporting a single source of truth across your organization. While metric views already existed in Databricks, the DAIS 2026 announcements demonstrate that these are becoming a much more mature product. For example, metric views now (or will soon) support multi-fact relationships, level-of-detail calculations, parameterized metrics that adapt to runtime inputs, advanced windowing calculations (with easier period comparisons and the ability to define periods with custom calendars), and both unaggregated and aggregated materializations.
For further reading on how to enable an AI-ready semantic layer, check out our recent blog series:
- What is an AI-Ready Semantic Layer and What does it Unlock?
- From Concepts to Application – What Does an AI-ready Semantic Layer Look Like?
- Building AI-Ready Semantic Layers
- Unity Catalog Metric Views
Governance Layer
Agentic analytics introduces a fundamental tension: how do you enable flexible, autonomous systems while maintaining enterprise-grade governance? Databricks’ answer is the Unity AI Gateway, which provides centralized enforcement of policies and guardrails across LLMs, APIs, and MCP servers; fine-grained control over what agents can access and do; end-to-end tracking of requests, decisions, and actions for auditability. This ensures that as agents become more powerful and more embedded in organizations, they remain secure, compliant, and observable.
Architecture Layer
Perhaps the most underappreciated challenge to enabling agentic analytics is infrastructure. Most existing data platforms were not designed for a world where thousands of agents are simultaneously querying data, triggering jobs, and interacting with systems.
Databricks is addressing this through continued investment in the Lakehouse architecture, most notably through their new ultra-low-latency engine Lakehouse//RT.
Key Takeaways
Together, these announcements signal a cohesive platform strategy that enables analysts to seamlessly leverage dashboards, chat interfaces, and autonomous analytics systems. For the past decade, the focus has been on improving access to insights via dashboards. But as the Databricks Data + AI Summit 2026 made clear, the next wave is about embedding AI directly into workflows, decisions, and actions.
Organizations that succeed in this new paradigm need to invest in a trusted semantic layer to ensure consistency across humans and agents, establish clear governance frameworks to balance flexibility with control, understand which tasks agentic workflows are better suited for and when dashboards or deterministic pipelines should be used instead, and build for scale from day one, anticipating a world where agents are primary consumers of data.
At Aimpoint Digital, we’ve built our teams around exactly this kind of end-to-end journey. If you’re ready to move beyond the dashboard and put agentic analytics to work in your organization, connect with our team to find the right starting point for your business.
.png)