News from the Databricks ecosystem.
Posts from databricks.com, MLflow, and dbt Labs — three trusted sources covering the platform, the open-source projects around it, and the data tooling layer most teams pair with it. Summarized for scanning.
This week
13 articlesAutomate Data & KPI Monitoring with SQL Alerts
In many organizations, data monitoring is still a manual, repetitive routine: open...
How to Build Real-Time Fraud Detection using Spark Real-Time Mode and Lakebase
Build real-time fraud detection with sub-second intervention using Spark Real-Time Mode and Lakebase. This unified platform processes high-throughput data streams, executes low-latency ML models, and serves explainable fraud scores to reduce detection lag and operational complexity.
How Databricks Genie improves supply chain visibility with real-time AI analytics
Databricks Genie now offers a conversational AI layer for real-time production intelligence, giving manufacturing VPs of Operations direct access to unified operational data. This eliminates data access bottlenecks, allowing leaders to ask context-aware questions and accelerate decision-making without complex analyst requests or BI tool training.
A CFO’s guide to managing value-based care financial performance
Databricks Genie for Value-Based Financial Intelligence helps CFOs bridge the VBC Financial Intelligence Gap by enabling conversational querying of integrated clinical-financial data. This allows healthcare finance leaders to quickly monitor patient utilization, track per-member-per-month costs, and identify clinical outliers essential for managing value-based care financial performance.
Stop rogue AI: How Unity Catalog secures your agent actions
The risks of agentic AI are no longer theoretical. Agents connected to external tools...
Why AI Security Infrastructure is Now a CMO Priority
Databricks launched Lakewatch, an open, agentic SIEM built on the Lakehouse, to bring security detection to where enterprise data already lives. CMOs must now prioritize AI security infrastructure and engage in enterprise data and platform decisions to scale AI safely, as marketing teams are increasingly targeted by agentic cyberattacks.
Databricks context engineer associate: the industry’s first certification for reliable AI agent systems
Databricks launched the industry's first certification for reliable AI agent systems: the Databricks Context Engineer Associate. This new certification validates skills in designing, managing, and governing AI context effectively.
Introducing AI spend controls with Unity AI Gateway
Unity AI Gateway now includes AI Spend Controls, offering proactive budget alerts across users, workspaces, and accounts to monitor and contain AI costs. These controls integrate with Unity Catalog system tables and Databricks budgets for unified governance of AI usage, cost visibility, and operational accountability.
How to safeguard AI workloads with Unity AI Gateway Guardrails
Unity AI Gateway now ships with pre-built and custom guardrails to protect sensitive information and ensure secure, compliant AI outputs. These guardrails are integrated with the Databricks lakehouse for simplified observability, monitoring, and evaluation of your AI workloads.
What’s new in Unity AI Gateway: service policies, guardrails, observability, and cost controls for AI agents and MCPs
Unity AI Gateway now includes service policies, guardrails, observability, and cost controls for AI agents and MCPs. This enables Databricks practitioners to govern model calls and agent actions with consistent policies, full visibility, and enforceable cost controls for production AI.
How Deutsche Börse built a generative AI tool to tackle the large-scale migration of Zeppelin notebooks to Databricks
Deutsche Börse built a Databricks App to automate the migration of Zeppelin notebooks, reducing redevelopment time from hours to minutes. This solution tackles a large-scale migration challenge by handling structural conversion and generating AI-assisted prompts for reconstructing notebook logic.
Announcing the Databricks analytics engineer learning pathway
A new learning pathway for Databricks SQL practitioners is now available on Databricks Academy, covering skills to use the full SQL ETL toolkit for data modeling, pipelines, semantic layers, and conversational agents. Courses are offered in self-paced and instructor-led formats, and are included with any active Databricks Learning Subscription.
The question your commercial data should already be able to answer
Databricks and Veeva now embed Genie AI agents and AI/BI dashboards directly into Veeva Vault CRM, enabling life sciences commercial teams to get real-time answers to their questions without leaving their workflow. This unified Databricks lakehouse with Unity Catalog delivers governed commercial data to every persona, from sales reps to MSLs, in the format and depth their role requires.
Last week
17 articlesPipelineIQ: Forward‑Looking Sales Intelligence That Drives Action
PipelineIQ is a new AI solution that provides prescriptive "Next Best Actions" for sales reps and managers, shifting focus from retrospective forecasting to immediate, forward-looking guidance. It's built to work with imperfect CRM data, extracting signals like champion strength and procurement stalls to deliver clear outcomes: Walk, Pivot, or Accelerate for every opportunity.
Backstage with Lakebase, part 2
Lakebase enables running production OLTP applications like Backstage on a serverless Postgres surface within Databricks, offering 1-second database branching and sub-4-second point-in-time recovery for schema migrations. Unity Catalog unifies governance for operational databases, providing single SQL query auditing, automatic row-level security propagation to branches, and zero-ETL cost attribution for FinOps.
Expanded interoperability with Unity Catalog Open APIs
Unity Catalog Open APIs now offer expanded interoperability, with external access to UC managed Delta tables in Beta and credential vending generally available with M2M OAuth support. External engines like Apache Spark, Flink, and DuckDB can now create, read, and write to UC managed Delta tables, leveraging Delta Lake's new catalog commits feature for safe concurrent writes and audibility.
From manual to autonomous: how AI agents are transforming electric grid operations
AI agents are transforming electric grid operations from manual to autonomous control, addressing unprecedented demand and complexity. Hawaiian Electric saw a 60X reduction in document query times, from five minutes to five seconds, in just two weeks with early AI agent deployment.
Data quality is the AI strategy
Your AI strategy hinges on data quality, starting with fixes to transactional systems. Organizations prioritizing value creation with unified data will benefit most, as tools and models constantly evolve.
The Rosetta stone of CPS: Claroty’s AI-powered library
Claroty's AI-powered CPS Library, built on Databricks Custom Agents and Delta Lake, automates entity resolution for 17M+ industrial and healthcare assets, solving the asset identity crisis where 88% of CPS devices lack exact product codes. This multi-agent AI system improves vulnerability attribution accuracy by over 25% and provides new security recommendations for over 56% of analyzed devices.
Clinical operations intelligence belongs on the Lakehouse
The Site Feasibility Workbench, an open-source Databricks App, now enables clinical trial site selection entirely within the Databricks workspace, eliminating external API calls and synchronization pipelines. This solution addresses the architectural challenge of disconnected clinical operations data, improving enrollment target attainment with TA-segmented LightGBM models and auditable SHAP-driven explanations.
ABAC row filtering and column masking policies, governed tags, and data classification are now generally available in Unity Catalog
ABAC row filtering and column masking policies, governed tags, and data classification are now generally available in Unity Catalog. These capabilities unify data governance, eliminating manual security and ensuring consistent, real-time protection across your data estate.
The Rise of Sports Intelligence: How the Lakehouse Turns Tracking Data into Competitive Advantage
Pro teams now leverage the Lakehouse to transform exploding tracking and biomechanical data into sports intelligence, driving real-time decisions on the court, in training, and in the front office. The Databricks Data Intelligence Platform acts as the governed "sports brain," unifying diverse data with Lakeflow, Unity Catalog, ML, and AI Search to power proactive injury management, coaching insights, and next-gen fan experiences.
How CFOs in consulting can recover margin with Databricks
CFOs in consulting can recover margin by unifying finance data and automating workflows on Databricks. Early adopters are seeing tens of millions in DSO optimized, 80% faster reporting, and 50%+ faster close cycles.
Announcing Native Lakehouse Sync
Native Lakehouse Sync (Public Preview) now automatically replicates Lakebase Postgres data into Unity Catalog managed tables, eliminating pipelines and external compute. This enables live ML features, operational data as the Bronze layer with full SCD Type 2 history, and built-in audit capture, all with zero Postgres performance impact and no added cost.
Announcing Databricks Student Fellows
The Databricks Student Fellows Program is now live, offering students an exclusive opportunity to develop in-demand data, AI, and computer science skills. Fellows will gain hands-on experience with the Databricks platform, earn certifications, build projects, and unlock career pathways.
The Convergence of Open Table Formats and Open Catalogs: Catalog Commits is Generally Available
Catalog Commits take a big step forward in unifying the lakehouse by aligning Delta...
Faster Queries and New Capabilities with the Open-Source Databricks JDBC Driver
The new open-source Databricks JDBC driver delivers up to 30% faster large result retrieval and adds support for multi-statement transactions, stored procedures, Arrow compatibility, and Unity Catalog metric views. This fully owned, open-source driver enables faster fixes, external contributions, and tighter platform integration.
Unlocking the Archives: Turning Unstructured Documents into a Searchable Database for Groundwater Discovery
MapAid partnered with Databricks for Good to transform 700 scanned hydrogeological documents into a searchable database using multimodal AI. This serverless pipeline classifies documents and extracts water-related information, enabling researchers to quickly find historical studies and well records for improved groundwater prediction.
Predictive Quality Starts Where Defect Detection Stops
Databricks Genie now enables quality leaders to interrogate their full operational dataset using natural language, synthesizing data from inspection and supplier lots in a single query. This eliminates data latency from fragmented systems, shifting quality from reactive documentation to predictive intervention for reduced scrap and improved margins.
Retail markdown optimization: from reactive markdowns to proactive
Databricks Genie for Merchandise Intelligence now enables retail CMOs to move from reactive to proactive markdown optimization. It provides instant, natural language access to synthesize critical data like trends, inventory, and pricing, allowing for earlier trend detection and improved margins.
Week of May 4
23 articlesHow Superhuman and Databricks built a 200K QPS inference platform together
Superhuman migrated their 200K QPS custom LLM inference to Databricks FMAPI Provisioned Throughput, achieving sub-second P99 latency and offloading infrastructure management. Joint engineering delivered 60% per-GPU throughput gains and reduced serving costs through FP8 quantization and Hopper architecture optimizations.
Using MemAlign to Improve Evaluation of Traditional Machine Learning in Genie Code
MemAlign, an open-source MLflow framework, significantly improved the evaluation of traditional machine learning in Genie Code by reducing LLM judge error by 74-89% on key dimensions. This alignment was achieved with ~50 labeled examples, demonstrating the importance of both semantic and episodic memory for closing the gap between LLM judges and human experts.
Addressing HR's widening capacity gap with AI
HR's capacity gap is costing millions, but AI adoption can help. Databricks' lakehouse architecture paired with MathCo's NucliOS platform provides the secure, scalable foundation for phased AI transformation, from data consolidation to AI-powered workflows.
MCP Marketplace Brings Real-Time Intelligence to Agentic Applications
MCP Marketplace now provides real-time external intelligence for agentic applications, with partners like You.com and Moody's offering governed data. Lakebase and Genie enable end-to-end workflows, allowing agents to maintain context and surface decisions to business users for review.
Pushing the Frontier for Data Agents with Genie
Genie is Databricks’ state-of-the-art data agent designed for answering complex questions...
Energy trading analytics in a real-time market
Databricks Genie now provides energy traders and portfolio managers instant, conversational access to critical trading data, eliminating the structural revenue problem caused by analytical lags in a real-time market. This enables optimal decision-making in a highly volatile, data-intensive environment with 15-minute price changes.
Operating room utilization is hiding in your scheduling data
Databricks Genie for Surgical Operations Intelligence is now available to help healthcare operations leaders close the "Operational Intelligence Gap" by enabling real-time, conversational querying of surgical scheduling data. This allows for immediate intervention on OR utilization, shifting capacity management from reactive to proactive and addressing forgone revenue and unmet patient needs.
Why telecom churn prediction misses the intervention window
Databricks Genie for Retention Intelligence helps telecom leaders act on early churn signals by providing real-time, natural language query access to customer data. This enables timely interventions, addressing the "Velocity Problem" where traditional churn models often identify at-risk customers too late.
Growth Analytics Is What Comes After Growth Hacking
Databricks AI/BI Genie resolves the fragmented data architecture bottleneck for growth analytics, enabling leaders to conversationally interrogate unified acquisition, behavioral, and revenue data. This provides a structural competitive advantage through faster spend reallocation and learning, moving beyond growth hacking to data-driven precision.
Real-world evidence for medical affairs: who can actually use it?
Databricks Genie for Medical Affairs Intelligence helps close the RWE Fluency Gap by enabling conversational querying of complex scientific questions against RWE data. Medical Affairs leaders can now generate insights in seconds, a task that previously took data scientists days.
Wealth advisor productivity starts with the client conversation
Databricks Genie for Wealth Management Intelligence is now available, enabling wealth advisors to conversationally query client data for instant, contextual answers and remove information preparation burdens. This allows advisors to shift focus from information logistics to genuine insight during client portfolio reviews, improving productivity and the quality of client conversations.
How lakebase architecture delivers 5x faster Postgres writes
Lakebase architecture now delivers up to 5x faster Postgres write throughput for OLTP workloads by offloading crash-recovery tasks to distributed storage. This change reduces WAL traffic by 94% and improves read tail latency by 2x without compromising durability.
Why Talent Transformation Is the Missing Focus of Enterprise AI
Databricks Academy Pro is now available, offering unlimited self-paced, live, and hands-on training to accelerate enterprise AI adoption. Continuous workforce upskilling is critical for competitive advantage, and this new offering supports that journey.
Public health intelligence shouldn't require a data scientist
Databricks Genie now enables natural language access across public health datasets, delivering rapid, governed insights to improve decision-making and resource allocation. This means public health intelligence no longer requires a data scientist, accelerating real-time outbreak response and intervention.
Mean time to detect is a data access problem
Databricks Genie within Lakewatch enables natural language, agent-driven investigations to accelerate detection and response. This addresses the core issue of cross-system data integration that limits security analysts' investigation speed and effectiveness.
First-party audience data is the ad sales relationship now
Media companies can now leverage Databricks Genie for Ad Revenue Intelligence to close the "Insight Gap" and conversationally query their first-party data for audience insights. This enables VPs of Advertising Revenue to quickly understand audience composition and demonstrate post-campaign performance, crucial for winning RFPs against platforms with richer data.
Rethinking Distributed Systems for Serverless Performance and Reliability
Databricks' serverless compute required rethinking distributed systems to eliminate user-managed infrastructure and improve stability. Architectural innovations like separating applications from compute and intelligent workload routing deliver more stable, predictable, and cost-efficient performance.
The AI Scaling Gap Hiding in Digital Native Companies
Digital natives lead on AI ambition but lag traditional industries in scaling it, according to new Economist data. Learn what's behind this AI scaling gap and why it matters.
10 trillion samples a day: Scaling beyond traditional monitoring infra at Databricks
Databricks now processes 10 trillion samples daily, scaling beyond traditional monitoring infrastructure by rearchitecting TSDB and aggregation layers with customized open-source solutions. A novel Lakehouse-based platform, Hydra, provides rich debugging capabilities for high-cardinality metrics at 50x cheaper storage.
AI success starts with clean data, not just better models
Databricks and Kraken are key to solving the foundational data challenge, providing unified, well-documented data essential for AI success. Organizations that pair this unified data with deep business context and a data-literate culture are pulling ahead, making data a business asset rather than just an IT platform.
How nOps Rebuilt Their Cloud Optimization Platform on Databricks Lakebase, and Why Other ISVs Should Too
nOps, a Databricks Built On partner managing over $4 billion in annual cloud spend,...
Peril Predicts: Precision Payouts for a Volatile World
Databricks now helps insurers operationalize parametric insurance workflows, enabling faster catastrophe payouts using objective event triggers. The Geospatial Lakehouse facilitates ingesting catastrophe data and analyzing exposure at scale, essential for reducing basis risk and defining accurate payout triggers with geospatial analytics and catastrophe modeling.
The foundation of AI scalability: one team, one platform, one operating model
Databricks now offers a unified platform and operating model to scale AI, addressing fragmentation with reusable accelerators and shared governance. This enables business teams to move 10x faster by fostering a culture of continuous learning and experimentation.
Week of Apr 27
33 articlesThe Federal Data Paradox: Rich in Data, Poor in Access
Databricks Genie enables federal agencies to overcome siloed, legacy data infrastructure by providing a natural language interface for governed, real-time data access. This empowers mission experts to make faster, evidence-based decisions without requiring a technologist for every routine task.
Driving Budapest Forward: How BKK Uses Databricks to Transform City Mobility
BKK now unifies city mobility data from buses, metros, trams, and shared vehicles on Databricks, creating a single view for analysts and decision-makers. This Lakehouse powers minute-level tracking, route performance analysis, and predictive modeling for operational efficiency and proactive urban planning.
LLM Vs AI: A Practical Guide to Differences, Use Cases, and Tools
LLMs are a subset of AI, and this guide clarifies their practical differences, use cases, and tools. Understand how LLMs fit into the broader AI landscape and what that means for your Databricks workflows.
Model Risk Governance Is Not the Same as Risk Intelligence
Databricks AI/BI Genie for Enterprise Risk Intelligence now enables conversational interrogation of governed risk data, providing instant, accurate answers for real-time risk management. This closes the intelligence gap for CROs who previously navigated complex model outputs and data systems to get specific answers on credit concentration or stress test sensitivity.
Generative AI for Business: A Complete Strategy and Implementation Guide
Generative AI is projected to add $2.6–4.4 trillion in annual economic value, with 75% concentrated in customer operations, marketing, software engineering, and R&D. Achieve durable business value by inventorying proprietary data, prioritizing high-impact pilots, and embedding AI into existing workflows with cross-functional teams and responsible AI practices like RAG.
Data Science vs Data Engineering: Choosing Analysis or Infrastructure
BI reporting bridges raw data and operational teams by collecting, analyzing, and presenting data in structured formats. Effective BI relies on clean, integrated data flowing through ETL pipelines into a central repository, supporting both managed and ad hoc reporting.
AI Applications: Tools, Use Cases, and Platforms
AI applications span four capability tiers, each with distinct data requirements and evaluation frameworks, and enterprise deployments often stall due to inadequate data infrastructure. Production-grade model development, from prompt engineering to pretraining, is increasingly accessible with open-source LLMs, but requires pre-built governance and monitoring infrastructure for successful deployment at scale.
MLOps vs DevOps: A Practical Guide for Data Scientists and IT Teams
MLOps extends DevOps by governing code, datasets, and model artifacts, adding Continuous Training pipelines to automatically retrain models when data drift exceeds thresholds. This guide details a three-layer model for successful MLOps, leveraging DevOps CI/CD, ML orchestrators, and unified monitoring to close the feedback loop.
Top Data Warehouse Tools For Modern Data Analytics
The lakehouse architecture is the modern standard for teams needing both analytics and AI, offering ACID-compliant reliability and open storage formats for SQL, streaming, ML, and AI on a single governed data foundation. Evaluate data warehouse tools across six dimensions, including query performance, scalability, and total cost of ownership, to avoid the hidden costs of maintaining separate systems.
Unlocking SAP Business Context in Databricks with Semantic Metadata Delta Sharing
SAP Business Data Cloud now automatically syncs semantic metadata, including descriptions and key relationships, into Unity Catalog, making SAP data instantly AI-ready and more discoverable. SAP PersonalData governance tags are also automatically available in Unity Catalog, enabling fine-grained access controls with ABAC.
The marketing activation gap has a fix: Databricks and Stitch partner to turn data infrastructure into marketing performance
Databricks and Stitch now partner to bridge the marketing activation gap, turning your data infrastructure into tangible marketing performance. Learn how Stitch connects your Databricks environment to marketing operations, with real-world examples from QSR, retail, and healthcare brands.
Alert Fatigue Is a Business Risk
Lakewatch and Databricks Genie unify data for agentic, machine-speed threat detection, triage, and response, directly addressing the business risk of alert fatigue in SOCs. This new approach helps overcome fragmented telemetry and legacy SIEM architectures that create signal-to-noise challenges and limit effective threat detection.
Backstage with Lakebase
For thirty years, the operational database and the analytical database have been...
Shipping Faster isn’t Learning Faster
Databricks AI/BI Genie for Product Intelligence ships to resolve the architectural bottleneck of slow data velocity in product organizations. VPs of Product can now instantly query complex behavioral data with conversational AI, eliminating weeks-long waits for analyst support and specialized skills.
Why Your OEE Dashboard Is Lying to You
Your OEE dashboard is likely misrepresenting operational reality because critical production data remains siloed and inaccessible. Databricks Genie provides a conversational AI layer over your unified data platform, making operational systems instantly answerable in natural language to empower real-time decision-making.
The Turbine That Tried to Tell You It Was Failing
Databricks Genie now offers a conversational AI layer for real-time operational metrics like OEE, directly accessible to VPs of Operations. This eliminates data access bottlenecks, enabling faster decision-making by shifting from reactive reports to proactive intelligence from SCADA and MES logs.
Predicting Readmissions Isn't Enough. Acting in Time Is.
Databricks Genie for Clinical Outcomes Intelligence now enables CMOs to conversationally query patient and outcomes data in natural language, providing immediate, governed insights to prevent predicted readmissions. This directly addresses the gap between readmission risk prediction and timely intervention by eliminating data request delays and matching clinical decision velocity.
Clinical Trials Run Longer Than They Have To. That's a Patient Problem
Databricks Genie for Clinical Trial Intelligence now enables clinical operations VPs to interrogate full trial data in natural language for instant answers. This allows earlier intervention on site performance issues, shortening trial timelines and improving patient access to treatment.
Network Quality Is a Revenue Problem, Not a Technical One
Databricks Genie for Network-Commercial Intelligence bridges the gap between network performance and commercial data, enabling Chief Network Officers to prioritize network quality decisions based on commercial impact. This helps reframe network quality as a revenue problem, not just a technical one, by connecting operational telemetry to commercial context like SLA exposure and churn propensity.
Approximate Answers, Exact Decisions: New Sketch Functions for Analytics
Databricks now offers new sketch functions for approximate answers to analytics questions, including KLL quantile sketches for percentiles, Theta and Tuple sketches for audience overlap, and approximate top-K functions for real-time trending. These functions enable faster, more memory-efficient computations over massive datasets, with mergeable sketches for incremental updates and combined counting and aggregation.
Companies Winning with AI Built the Data Layer First
Companies winning with AI, like Trinity improving delivery by 15% and ETA models by 50%, built a unified, governed, and accessible data layer first. Consolidating fragmented systems into a single architecture enables real-time AI, faster decisions, and lower costs.
Rethinking SQL ETL for modern data platforms
Databricks now offers a unified platform for all SQL ETL, removing coordination overhead and letting teams ship faster on one governed system. This approach addresses the fragmented SQL ETL that drives hidden cost, brittle pipelines, and slow incident resolution across warehouses, orchestrators, and tools.
Stripe data now available on Databricks via Databricks Marketplace
Stripe data is now available on Databricks Marketplace, enabling you to activate a Stripe data pipeline with Delta Sharing in minutes and instantly power AI applications. Share Stripe payment and business data directly into Unity Catalog to create a single source of truth and query live payment data for models, agents, and Genie workspaces.
Databricks and Stripe Projects: Infrastructure Built for Agents
Stripe Projects, a new agent-first CLI, now lets AI coding agents discover, provision, and pay for Neon Postgres databases directly with no human-in-the-loop. Databricks is a launch partner, enabling agents to spin up production-ready Postgres databases in seconds, backed by Lakebase's serverless architecture.
Agents are ready but your architecture probably isn't
Agentic systems are failing in production due to siloed data, poor governance, and analytics-focused infrastructure. CDOs and CTOs need a transactional database built for agents and a clear vision for success.
Interoperability Between Unity Catalog and Google BigQuery via Catalog Federation
Google Cloud now supports catalog federation to Unity Catalog, enabling BigQuery users to read tables in Unity Catalog without duplication. Unity Catalog also supports catalog federation to Google Cloud's Lakehouse, allowing it to read Iceberg tables written from BigQuery and other engines.
Built In, Not Bolted On: What AI-Native Actually Means in Cybersecurity
Databricks now offers a truly AI-native cybersecurity solution, architected with intelligence at its core and leveraging proprietary telemetry for a defensible advantage. This approach prioritizes defining shared outcomes for cross-functional alignment, rather than simply selecting shared tools.
Operationalizing AI for public sector fraud prevention
Databricks now offers a scalable approach to AI-powered fraud prevention for public sector agencies. Learn how clean data, automation, and real-time insights can be integrated into daily workflows to improve decision-making.
From months to minutes: Building real-time clinical data pipelines with natural language
Databricks and Redox now enable real-time clinical data pipelines from EHRs to Unity Catalog with natural language prompts, reducing integration time from months to minutes. This partnership allows AI outputs to be written back into the EHR in real time, transforming Databricks into an operational layer for point-of-care interventions.
Agentic Data Engineering with Genie Code and Lakeflow
Genie Code, an autonomous AI partner for data engineers, is now integrated directly into Lakeflow. Data engineers can leverage Genie Code within Lakeflow's Pipeline Editor and Jobs for the full data engineering lifecycle, from development and orchestration to monitoring and debugging.
Securely send first-party conversion signals with Snapchat Conversions API on Databricks Marketplace
Snapchat Conversions API is now on Databricks Marketplace, letting you send web, app, and offline events from your Lakehouse directly to Snapchat. Improve Event Match Quality and campaign measurement by deploying a pre-built notebook for governed, server-side event delivery.
How leading tech companies are killing the builder’s tax with Lakebase
Databricks Lakebase is helping leading tech companies eliminate ETL and reverse ETL by unifying operational and analytical data. This enables real-time intelligence for apps and AI systems, operating directly on fresh, low-latency data.
Inside one of the first production deployments of Lakebase: LangGuard's agentic workflow governance engine
LangGuard's agentic workflow governance engine, one of the first production deployments of Lakebase, extends Unity Catalog and AI Gateway with runtime enforcement for autonomous AI agents. Lakebase provides the elastic, low-latency operational data layer for LangGuard's GRAIL™ data fabric, enabling real-time policy evaluation without impacting agent performance.
Week of Apr 20
11 articlesThe next generation of Databricks Genie
Genie now answers questions beyond Genie Spaces, connecting to external knowledge stores like Google Drive and SharePoint. This next generation of Genie, previously Databricks One, is available on web and native mobile apps.
Model Risk Management in 2026: A Banker’s Guide to the Revised Interagency Guidance
The new interagency guidance for model risk management, effective April 17, 2026, shifts to a risk-based, principles-driven framework, demanding a unified, governed lifecycle for all models, including GenAI, with evidence of governance generated automatically. Databricks offers a reference architecture to meet these revised expectations, integrating model development and validation into a single, auditable process.
OpenAI GPT-5.5 + Codex, now available and fully-governed in Databricks
GPT-5.5 and Codex are now natively available in Databricks, fully governed by Unity AI Gateway for permissions, cost controls, guardrails, and observability. This enables agent building with GPT-5.5 and natural language querying of enterprise data via Genie.
Operational databases: How they work and when to use them
Databricks is introducing the "Lakebase," a new open architecture combining transactional database speed with data lake flexibility and economics, designed to overcome the limitations of traditional operational databases for modern unstructured data and AI workloads. This allows for real-time processing and concurrent transactions directly on the data lake, eliminating slow ETL pipelines and supporting diverse data types.
Databricks partners with OpenAI on GPT-5.5
GPT-5.5 and Codex are coming soon to Databricks, governed by Unity AI Gateway, and cut OfficeQA Pro errors nearly in half. This partnership with OpenAI brings advanced models directly to Databricks users.
Announcing the Public Preview of Lakeflow Designer
Lakeflow Designer is now in Public Preview, offering a visual, no-code, AI-native interface for data preparation and analysis directly within Databricks. It leverages Unity Catalog for governance and generates production-ready code, providing step-by-step data previews for easier review of AI-generated transformations.
Are LLM agents good at join order optimization?
LLM agents can improve Databricks join order optimization, achieving 1.3x latency reduction in 80% of cases by reasoning through runtime statistics. This prototype demonstrates LLM agents' potential to act as data-driven DBAs, addressing cardinality misestimation challenges in complex SQL queries.
How conversational analytics removes the BI bottleneck
Databricks Genie and Lakebase are transforming BI by enabling conversational analytics with enterprise context, providing actionable insights beyond traditional dashboards. Operationalizing trusted AI-powered analytics, built on robust governance and semantic layers, is now crucial to avoid a competitive gap.
How to transform document activation workflows with Genie and Agent Bricks
Databricks shipped a solution combining AI/BI Genie, Agent Bricks, and Unity Catalog to automate document activation workflows. This enables multi-agent orchestration for extracting, processing, and activating data from diverse documents, improving efficiency and governance.
Beyond the spreadsheet: how Databricks is delivering the modern CFO in Financial Services
Databricks now offers a unified architecture for Financial Services CFOs, integrating real-time data, AI modeling, and governance to eliminate data fragmentation and slow reporting. This enables a shift from reactive reporting to strategic finance, with benefits like drastically reduced regulatory reporting times and AI-powered natural language querying of complex financial data.
AI App Development: Guide To Building AI-Powered Apps
Databricks Apps and Lakebase are purpose-built platforms that streamline AI app development by eliminating infrastructure, authentication, and data synchronization overhead. A structured process covering model strategy, prompt design, agent orchestration, and data prep, combined with rigorous quality gates, ensures production-grade AI applications.