Google Cloud expands AI agent observability and governance across BigQuery and Looker
·2026.11.14 02:17
Key point
The new framework captures telemetry from first-party and custom agents into BigQuery, enabling teams to track reasoning, control token costs, and evaluate accuracy before deployment.
Details
- Google Cloud announces expanded AI agent observability and evaluation capabilities across Google Data Cloud, providing unified visibility for first-party agents in BigQuery and Looker and custom third-party agents built with Agent Development Kit (ADK) or LangGraph.
- A unified telemetry framework captures agent data into customer-owned BigQuery tables, supporting out-of-the-box logging for Conversational Analytics agents and single-line SDK instrumentation for custom agents.
- Governance features include always-on admin audit logs, VPC Service Controls, customer-managed encryption keys (CMEK), and regional data residency, with privacy guardrails that skip logging when row-level security or dynamic data masking is active.
- FinOps controls allow teams to track granular token counts and latency, configure per-query BigQuery byte scan caps to block oversized queries, and enforce daily input and output token quotas (in preview) per project, user, or agent.
- Evaluation tools include Prism, an open-source tool for benchmarking agent accuracy using reusable test suites, and native BigQuery evaluations (in preview) that stream results into day-partitioned tables for side-by-side analysis with production logs.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.