At Data + AI Summit 2026, we're announcing a wave of new capabilities that bring the combination of context and control to the agentic era. In order to transition enterprises from narrow experimental AI pilots to production-grade automated workflows, we are expanding the Azure Databricks platform across four foundational pillars: establishing an ultra-fast, zero-copy real-time foundation with Agentic Data; embedding data-smart AI coworkers directly into daily productivity tools with Agentic Dev & Work; deploying autonomous, lakehouse-embedded personalization with Agentic Marketing; and anchoring the entire ecosystem under an intelligent, secure governance framework. Together, these advancements deliver a unified architecture designed to help your data, your teams, and your autonomous agents operate seamlessly natively on Azure.
To fuel autonomous agents with real-time data without forcing data replication into costly operational side-stacks, Azure Databricks introduces the first true LTAP (Lake Transactional/Analytical Processing) Architecture. This unified storage layer brings your analytical data, streaming pipelines, and live application transactions together into a single, shared copy of storage directly on the lakehouse.
As the transactional engine of this framework, Azure Databricks Lakebase delivers a fully-managed, serverless Postgres database purpose-built for the agent era. Featuring decoupled compute and storage, Azure Databricks Lakebase supports instant copy-on-write database branching to completely eliminate compliance risks when debugging production AI agents. Developers can spin up a full-fidelity branch of a live production database in seconds, allowing engineers to point GitHub Copilot agent mode directly at the temporary branch to safely reproduce edge cases, identify root causes, and deploy fixes through standard Git-based workflows.
For downstream analytical serving, Lakehouse//RT shatters the legacy scale-latency tradeoff. Powered by the vectorized Reyden engine, it delivers sub-second, millisecond-level response times for high-concurrency workloads directly on your data lake, creating an ultra-fast foundation that integrates seamlessly with operational dashboards and Power BI.
Lakehouse//RT ran more than a third faster on average than our prior warehouse on our healthcare dataset, with 10× faster queries. That translates directly to quicker information access and more decision time for our customers. We had considered a dedicated real-time system to augment our Lakehouse architecture, but Lakehouse//RT removed that need, giving us that speed natively with consistent governance.— Mehrshad Setayesh, SVP Engineering (Data, Platform, AI) at PointClickCare
Access any data stored in OneLake (Now Generally Available): Azure Databricks can query data stored in OneLake directly through Unity Catalog without copying data.
Store data in OneLake (Now in Public Beta): Azure Databricks can now store managed Delta tables natively in OneLake. Whether data is stored in OneLake or ADLS it is available zero-copy in OneLake for all Fabric engines.
The best AI insights are the ones that reach you without friction, which is why we’re bringing Genie natively into the collaboration tools where your teams already work and make decisions every day.
For teams working across the Microsoft ecosystem, that same data intelligence is now available directly within your everyday collaboration tools. Picture this: your VP of Sales pings you in Teams asking "What were our top accounts this quarter and why did we miss the Southeast target?" Instead of scrambling across dashboards and reports, you simply tag @Genie in the thread and your entire team gets a context-aware answer from your Azure Databricks lakehouse in seconds. Now in Beta, the Databricks Genie integration for Microsoft Teams and M365 Copilot extends AI-native intelligence across every chat and Copilot-powered workflow. Tap in Genie to answer that.
And available today, Databricks Genie works seamlessly with M365 Copilot Cowork. This integration will allow teams to anchor Cowork’s tasks with the Genie Ontology, bringing trusted data intelligence straight into their workflows.
Genie shifts analytics from a passive reporting dashboard to an active, data-smart AI coworker across your entire Microsoft surface area. This integration is fully governed by Unity Catalog, ensuring every answer is trusted, secure, and scoped to exactly what each user can see. Alongside this rollout, we are highlighting the complete Genie innovation framework:
For teams living in Excel, we’re meeting them where work already happens. The Azure Databricks Excel Add-in, now in public preview, brings your lakehouse directly into spreadsheets: no SQL, no per-user ODBC setup, and less friction.
With support for Unity Catalog metric views, data teams can define business logic once and make it instantly available in Excel and beyond, fully governed, secure, and consistent. And it’s not just read-only. The add-in also supports write-back, so users with permission can push updates from Excel straight into Databricks, closing the loop between analysis and action.
The result is faster, more reliable decisions by bringing governed lakehouse data and business logic directly to Excel users.
To further automate file processing across the entire enterprise ecosystem, the public Beta of the fully managed SharePoint Connector via Lakeflow Connect eliminates manual ingestion hurdles. This connector allows organizations to deploy automated, point-and-click ingestion pipelines for both structured sheets and unstructured files, such as PDFs, Word documents, and PowerPoints. By automatically streaming SharePoint file repositories directly into Delta tables, this integration ensures that downstream analytics pipelines, Genie One spaces, and Excel workbooks are constantly supplied with fresh, verified data without manual text extracts or risky file downloads.
To eliminate the operational complexity of siloed MarTech applications, we are introducing Azure Databricks CustomerLake: the industry's first Agentic Customer Data Platform (CDP) built natively inside the lakehouse foundation. Fully embedded within your secure storage boundary, CustomerLake equips data teams with autonomous Profile Agents to help transform raw data into business-ready Customer 360 profiles across fragmented sources. Simultaneously, a marketer-friendly workspace empowers business users with Campaign Agents to segment audiences, recommend next-best actions, activate across channels, and continuously optimize 1:1 personalized experiences.
What excites us most about CustomerLake and the new CDP capability is the ability to bring customer data together in a way that is actionable, timely, and scalable. By creating a more complete view of each customer, we can better understand behaviors, preferences, and needs across channels, which will help us deliver more personalized experiences and more relevant offers. Ultimately, we see this as a powerful step toward stronger engagement, deeper loyalty, and better outcomes for both our business and our customers.— Jay Malepati, Global Director, Customer and Marketing Data Science, Circle K
Powering these intelligent applications requires granular administrative control and semantic precision. The foundational intelligence layer of our platform is the Genie Ontology, a self-improving semantic context engine. Rather than requiring manual curation, the Genie Ontology automatically extracts table relationships, column metrics, and query popularity signals directly from your pipelines, eliminating AI hallucinations and ensuring that models accurately understand unique enterprise jargon.
To govern these models as they scale, the Unity AI Gateway serves as a centralized runtime registry inside Unity Catalog. It establishes strict, real-time rate limits, content filtering, and hard spend caps to guarantee predictable tokenomics across all automated workflows.
By connecting real-time data foundations directly to everyday tools like Microsoft Teams and Excel, Azure Databricks makes it simpler than ever to run and govern trusted AI workflows. Explore the updated product documentation or visit Databricks Academy to start putting these new capabilities to work today.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。