Databricks Enables CI/CD for AI/BI Dashboards via Asset Bundles
Databricks integrated AI/BI Dashboards with Asset Bundles to enable CI/CD workflows. Users can now manage dashboard lifecycles using YAML and Git. This move counters similar developer-first BI features in Snowflake and Microsoft Fabric. Data teams can now automate dashboard testing and deployment. This reduces production breakages and ensures consistent metrics across environments.
Insight Toast Expands Workforce as Block Automates 4,000 Roles With AI
Toast added 800 employees in 2025, a 14% increase that defies the fintech layoff trend. While Block cut 4,000 workers to pivot toward AI-led operations, Toast is using AI to augment its growing human workforce. Tools like Toast IQ now automate inventory and margin analysis. This highlights a strategic split between total automation and human-centric scaling.
Databricks Builds Reffy AI Assistant Using Mosaic AI Framework
• Lakebase Persistence
• Gemini 2.5 Scoring
• DSPy Agent Logic
LinkedIn & Medium
Internal Slides/Docs
Public Case Studies
Databricks launched Reffy, an internal RAG-based assistant indexing 2,400+ customer stories. It uses the Mosaic AI Agent Framework and Gemini 2.5 to automate knowledge retrieval. This implementation competes with enterprise search platforms like Glean and AWS Q. By centralizing tribal knowledge, Databricks reduced manual search time for 1,800 employees, proving the utility of custom-built internal agents.
SAP Restructures Board to Bridge Enterprise AI Execution Gap
SAP appointed Thomas Saueressig as Chief Customer Officer to lead the new Customer Value Group. This unit merges sales and delivery to accelerate AI adoption and cloud renewals. The move addresses a market-wide execution gap where 95% of AI pilots fail to scale. SAP is now aligning its entire customer journey to prioritize AI-driven business outcomes.
Microsoft Launches Season 2 of JavaScript AI Build-a-thon
• Advanced RAG Deep Dives
• Agent Builder (AITK)
• 6 Award Categories
• Product Demo Series
Microsoft launched Season 2 of its JavaScript AI Build-a-thon. The month-long event teaches developers to build local AI and multi-agentic systems using TypeScript. This arrives as TypeScript surpasses Python in GitHub popularity. Microsoft is pushing AI development into the web ecosystem. It focuses on autonomous agents and the Model Context Protocol instead of simple chatbots.
AWS Launches Nova Forge for Specialized Model Fine-Tuning
AWS launched Nova Forge to automate data mixing during model fine-tuning on Bedrock. It blends user datasets with foundation data to preserve general reasoning while adding domain expertise. This addresses catastrophic forgetting, a common failure in OpenAI and Google fine-tuning workflows. Developers can now build specialized agents that retain core logic without manual data balancing.
Google Makes AI-Powered Non-Skippable Video Ads Generally Available
Google launched non-skippable ad support for Video Reach Campaigns globally. The system uses AI to mix 6, 15, and 30-second formats across YouTube and Connected TV. YouTube currently leads the U.S. CTV market with a 12% share. This update automates brand awareness, competing with Meta’s Advantage+ by removing manual format testing and improving reach efficiency.