BigQuery chatbot connection and query preflight Reviewed: 2026-11-09 Editorial worksheet; not an AYD import or automatic cost control. No cloud job, customer dataset or AI model was tested for this worksheet. Never record credential contents, private keys or sensitive result rows here. Workflow and product/version: Data owner and approved test scope: Execution principal (non-secret identifier, kept in your internal copy): Job project: Fully qualified source objects: Dataset region or multi-region: Metadata discovery result: Actual application execution result: Required audience/row/column restrictions and enforcement path: Reviewed read-only SQL file: reviewed-query.sql SQL revision and reporting period: Dry-run environment and identity match confirmed: Command reference (replace placeholders; requires your approved CLI setup): bq --project_id=JOB_PROJECT --location=DATASET_LOCATION query \ --use_legacy_sql=false --dry_run < reviewed-query.sql Validation result: Estimated bytes and any estimation limitations: Billing model (on-demand/capacity/other): Agreed per-job limit, where applicable: Evidence the actual submitting path enforces that limit: Project/user quota configuration and owner, if applicable: Follow-up/retry controls and evidence: Do not infer application limits from a separate CLI/console test. Known-answer question and independently checked reference: Observed rows/aggregate compared with reference: Allowed/denied access tests and observed results: Unresolved error (redacted): Next change, owner and recheck trigger: Go/no-go and reason: Fictional decision example only: 80 GiB estimate vs 2 GiB approved limit requires review, not an automatic limit increase. No measured cost is implied.