Industry story
Google BigQuery Conversational Analytics reaches general availability
ai-in-adtech engineering measurement
Google making Conversational Analytics in BigQuery generally available is less about SQL-free queries and more about whether a Gemini-powered agent can run a full analytical investigation autonomously — build the plan, execute it, schedule it, surface anomalies daily — without a data engineer in the loop. That's the actual bet here. The cross-cloud reach (Databricks, AWS Glue, SAP, Salesforce) and EU/US data residency commitments are the enterprise checklist items that get this past procurement; the agentic deep-dive mode is what determines whether it changes how analysts actually work. Watch whether "scheduled agentic workflows" holds up under real-world data messiness — that's where the GA label either earns its keep or becomes a roadmap item.
Full analysis
Google Cloud has moved Conversational Analytics in BigQuery to general availability (GA), enabling both technical and business users to query data, run multi-step analyses, and generate visual reports using plain-language prompts — no SQL required. The feature is powered by Google's Gemini models (Google's family of large language models) and is built directly into BigQuery Studio with no setup needed. It supports cross-cloud data sources including Databricks Unity, AWS Glue, SAP, and Salesforce, and now guarantees EU/US data residency for both stored data and ML processing.
The GA release introduces agentic (autonomous, multi-step) capabilities: a 'deep-dive mode' where the agent autonomously builds an analytical plan and runs a full investigation, and scheduled agentic workflows that monitor data and deliver proactive reports (e.g., a daily anomaly detection summary). Trust and explainability features include visible reasoning steps, SQL inspection before execution, source citations, and long-term memory so users don't have to repeat context. Governance controls inherit BigQuery's existing permissions model, with added support for customer-managed encryption keys, VPC service controls, and per-agent query-size caps.
Comments