Co-ordinating AI/BI Genie Spaces with Databricks Agent Bricks
Description
One of the most exciting, impressive announcements coming from the Data + AI Summit 2025 was this new Agent Bricks thing - but what can you actually use it for? How do you get started? What actually is it? In this video, Simon walks through a very first use case for Agent Bricks - building a co-ordinating agent that can sit across multiple Genie Spaces and answer questions accordingly! Pretty damn cool - it's currently in Beta, so don't get too exciting about automating the entirety of your production jobs just yet, but it's a VERY exciting glimpse at how we're going to solve a whole bunch of upcoming problems!
Description from YouTube. Full content on the video page.
More from Advancing Analytics
NewsEp 6: All About Genie App Builder | James Bentley
The video teaches how the Databricks Genie App Builder allows users to create governed, enterprise-ready applications using plain-language descriptions tied directly to Unity Catalog data. It demonstrates how features like app spaces and serverless micro apps enable scalable development while reducing infrastructure and governance challenges.
NewsEp 5: All About Genie One & Genie Ontology | Ashley Warren
Databricks announced Genie Ontology, a knowledge graph that scans business data and documents to understand organizational context, enabling Genie One—an AI co-worker deployed in Slack and Teams—to answer data questions with 84.5% first-attempt accuracy while taking actions like scheduling tasks and creating reports. The system shifts analyst responsibilities from manual querying to data governance and quality management, with trust and metadata documentation becoming critical for accuracy.
NewsEp 4: All About Genie ZeroOps | George Dutton
The video introduces Genie Zero Ops, an autonomous background agent designed to monitor data and AI assets within Databricks and resolve pipeline or model issues. It details how the agent traces root causes, tests fixes safely in zero-copy sandbox clones, and presents approved solutions via an inbox-style interface.
EventsEp 3: All about Lakehouse//RT and LTAP | Peter Sach
Databricks has announced Lakehouse RT, a real-time data warehouse engine powered by the new Raiden engine that delivers millisecond-level query performance directly on existing lakehouse tables without data replication. The company also introduced Lake Transactional and Analytical Processing (LTAP), which integrates the Postgres-based Lakehouse with Unity Catalog to enable seamless operational and analytical workloads on a single copy of data.
NewsGenie Code - Unleashing Agent Mode!
Databricks has rebranded its Databricks Assistant as Genie Code and introduced a new agent mode. This new feature allows the assistant to build entire systems, including notebooks, tests, jobs, and dashboards, from a single prompt rather than just suggesting code for individual cells.
EventsFive Hours of DAIS 2026 Announcements in 30 Minutes!
The video summarizes key announcements from the Databricks Data and AI Summit, focusing on new "Genie" branded AI tools for business users and developers, and significant engineering updates like the rewritten Spark engine "Raiden" and self-healing pipelines "Zero-Ops." It also introduces new Databricks SaaS applications like Lakehouse for cybersecurity and Customer Lake for CDP.