BigQuery is the catch-all source: if a number lives in your warehouse, Jents can measure how your agents move it.
What you can measure
Pick any numeric column across your datasets and Jents will count, sum, average, min or max it over a time window, or build a ratio of two (win rate, average order value, cost per lead). Jents auto-classifies a column as money, time, percent or a plain count from its name, so the dollar math is set up for you.
Good things to measure: revenue, orders, signups, activated accounts, tickets resolved, hours saved.
Connect it
In Google Cloud, create a service account with the BigQuery Data Viewer + Job User roles (read-only), then generate a JSON key.
In Jents, Settings → Integrations → BigQuery → Connect → paste your project ID and the service-account JSON. Jents verifies it can list a dataset, then encrypts it.
What Jents reads
Only scoped aggregate queries, a single number per read (for example, the SUM of a revenue column between two dates) plus a row count for confidence. It never reads rows or record contents.
Specs
| Spec | Detail |
|---|---|
| Auth | Service-account JSON (read-only) |
| API | BigQuery REST v2 (standard SQL) |
| Access | Read-only, scoped to aggregate queries |
| Reads | One server-side aggregate per metric |
| Refresh | Daily, plus on-demand |
Privacy & security
Read-only, numbers only, never rows, names, or record content. Every read goes through one audited, aggregate-only path and lands in your audit log. Credentials are encrypted at rest, we never train on or share your data, and you can disconnect anytime, see Privacy and Trust.