Data Analytics News: Databricks Genie Boosts 91% Score

This week’s data analytics news is dominated by agents. Databricks rebuilt its Genie assistant as an always-on autonomous worker, Snowflake pushed its own background agents and a governance gateway, and Teradata showed what a legacy vendor’s cloud pivot looks like in the numbers. Below are the stories worth your attention, with the primary source for each. Nothing here is older than about ten days, and where a vendor’s claim is unverified we say so.

Data analytics news: Databricks Rebuilds Genie as an Always-On Agent

On November 11, Databricks launched a second-generation Genie that works in the background on its own compute: you give it a goal and boundaries, and it delegates to subagents, tracks progress and pings you when it needs a decision. Databricks says Genie with its Ontology layer answered 91% of Turing benchmark task runs correctly, 1.6x a stronger frontier agent while using under half the tokens. That is a vendor-reported result, so treat it as a claim to test on your own data.

Governance runs through Unity Catalog, with approval rules, budgets and per-user spending caps. The press release lists Genie and Genie Ontology as generally available, while the always-on mode and Genie Search are still “coming soon” to preview. More detail is in the Databricks engineering blog.

Snowflake Answers With Background Agents and a Cortex AI Gateway

At Expedition 2026 on November 3, Snowflake announced CoCo and CoWork updates: CoCo Automations, the Async Agent API and skills sharing are generally available, while always-on CoCo and CoWork Bots arrive first as a private preview. Snowflake says CoWork is used by more than 5,800 accounts and CoCo by more than 9,100.

A second release introduced the Cortex AI Gateway and Business Ontology, a single control plane for agent model access, MCP tool calls, cost governance and observability. The gateway is generally available, but several of its safety features are still in preview. Snowflake cites a sponsored MIT Technology Review Insights survey of 400 executives in which 55% blamed missing business context for poor agent insights.

Teradata Cloud ARR Climbs to $700M as Revenue Slips

Teradata’s third-quarter results filed on November 5 are a useful reality check on the shift to cloud. Total revenue was $399M, down 4% from $416M, but public cloud ARR rose 11% to $700M and total ARR reached $1.512B. The company raised its full-year earnings guidance and highlighted Teradata Factory, its integrated GPU system, plus an integration with Microsoft OneLake. If your team still runs on-premises warehouses, this is the migration math vendors are living through too. Cloud growth is real, but it has not yet offset the decline elsewhere.

EXL Buys Aimpoint Digital to Add 250 Data and AI Experts

Also on November 5, EXL announced it has acquired Aimpoint Digital, a data and AI consulting and platform engineering firm. The deal adds roughly 250 AI engineers, architects and strategists to EXL’s workforce of more than 74,000, and the terms were not disclosed. The move reflects a wider pattern: services firms are buying specialist data talent so they can staff the “forward deployed” agent projects that platforms like Genie and CoCo now make possible.

Ocient Raises $100M to Push Agentic Analytics

We covered the round in detail in our earlier Ocient Series C write-up, but the headline numbers are worth repeating. Blocks & Files reports a $100M Series C co-led by BOKA Capital and OCA Ventures, taking total funding to $260M, with the company expecting to triple revenue this year. Ocient says the money will extend OcientAIQ to more customers. It is one more sign that investors still back databases built for AI-scale analytics.

What This Data Analytics News Means for Your Stack

Three practical takeaways. First, “always-on” agents are arriving in preview, not production, so pilot them on low-risk questions and watch the spending caps. Second, governance is now the product: Unity Catalog, the Cortex AI Gateway and Business Ontology all exist to stop agents guessing at what your metrics mean. Third, none of this removes the need to write correct queries. Agents still generate SQL, and you cannot review what you do not understand, so brush up on filtering with the WHERE clause and the anatomy of a query in our SQL series. We will keep tracking this data analytics news as the previews turn into releases.