In partnership with
News of the day
1. Databricks introduces agent bricks, a governed enterprise agent platform for real business data and production-grade ai workflows. the platform unifies model access, execution, governance, and context, enabling teams to build and deploy agents that operate securely and reliably in production. → Read more
2. SAP integrates agentic AI into human capital management to streamline operations and reduce costs by automating troubleshooting and enhancing decision-making. → Read more
3. Kumo’s new foundation model, KumoRFM-2, replaces months of data science engineering with plain-English queries, scaling to over 500 billion rows of data. → Read more
4. Gemini robotics-er 1.6 enhances reasoning with unprecedented precision for real-world robotics tasks and safety improvements → Read more
Our take
Hi Dotikers!
Yesterday, the Stanford HAI report highlighted an uncomfortable paradox: model capabilities are soaring, but transparency around how they work is eroding. This week, Databricks shows up with a response that reads like an ad for the very problem Stanford just described.
Agent Bricks, Databricks' governed agent platform, has just hit several significant milestones.
Document Intelligence moves to general availability, enabling structured information extraction from PDFs, tables, and reports directly within enterprise data workflows. Custom Agents are also now generally available on Databricks Apps, giving engineering teams the ability to deploy their own agents using their preferred frameworks (LangChain, LangGraph, or LlamaIndex) without rewriting code or managing infrastructure. Agent Mode lands in Genie, Databricks' natural language query tool, effectively turning data analysis into agentic operations.
What sets Databricks apart in a market drowning in agentic promises is its positioning around governance. Unity Catalog, the permission management system that has covered structured data for years, now extends to agents, models, and MCP servers. In theory, every interaction is traced, every access is scoped to the actual user's permissions -- not the overly broad privileges of a service account. Franklin Templeton and AstraZeneca cite concrete results: regulatory analyses completed in seconds instead of days, extraction of 400,000 clinical documents without a single line of code.
In practice, governance remains the blind spot of most enterprise agentic deployments. Overly broad access, audit trails that vanish the moment an external agent enters the loop, permissions that drift between the data warehouse and the AI layer -- these are real problems that Databricks is trying to absorb into a single platform. It's ambitious, and not necessarily wrong. But any company that thinks a well-configured Agent Bricks setup spares them from thinking about their underlying data strategy is setting itself up for disappointment. The tool doesn't replace the plumbing.
Alex.
AI Agents Are Reading Your Docs. Are You Ready?
Last month, 48% of visitors to documentation sites across Mintlify were AI agents—not humans.
Claude Code, Cursor, and other coding agents are becoming the actual customers reading your docs. And they read everything.
This changes what good documentation means. Humans skim and forgive gaps. Agents methodically check every endpoint, read every guide, and compare you against alternatives with zero fatigue.
Your docs aren't just helping users anymore—they're your product's first interview with the machines deciding whether to recommend you.
That means:
→ Clear schema markup so agents can parse your content
→ Real benchmarks, not marketing fluff
→ Open endpoints agents can actually test
→ Honest comparisons that emphasize strengths without hype
In the agentic world, documentation becomes 10x more important. Companies that make their products machine-understandable will win distribution through AI.
Meme of the day
