# Looker can check six of the ten AI-readiness conditions with its own features

GUIDE, STACK, read 13 Nov 2026

**The short answer.** Google's Looker documentation covers six of the ten AI-readiness conditions natively: documented, certified, quality, usage, metric definition and access. Owner and fresh are partial, and sensitivity and lineage show none found on 13 Nov 2026 in the pages read. The checks live in LookML and the System Activity explores.

## Six conditions are native, two partial, two not found

We counted the ten conditions against the Google Cloud pages listed below, read on 13 Nov 2026. Six have a native Looker feature: documented, certified, quality, active use, metric definition and access. Owner and fresh are partial. Sensitivity classified and lineage show none found on 13 Nov 2026 in the pages read.

In Looker the metric definition and the access rule are both written in LookML, so the checks run against code that is reviewed in a pull request.

## LookML descriptions now feed the agent, so they are the documented check

Google's Conversational Analytics best-practice page says the agent reads "labels, descriptions, and synonyms that are defined in the LookML model". It also says "Conversational Analytics uses descriptions to better identify field meanings and map user terms." The check is to list the fields in each Explore and count the empty description and label parameters.

The description parameter exists for fields and for Explores. The tags parameter "lets you add text strings to a field", which is a place for a team convention, though Looker does not interpret the tag.

## Content certification and LookML data tests cover certified and quality

Per Google's content certification page, it "lets trusted users certify Looker dashboards, Looks, Explores, LookML dashboards, and self-service Explores" after a manual review. Looker keeps three trust levels, and an uncertified item carries an "ungoverned" warning badge. Check the badge on each Explore an agent will query. Certification is on by default.

For quality, the Assert Validator "runs the LookML data tests that were created by Looker developers in your LookML project". Check that every agent-facing Explore has data tests and that the latest run passes. The Content Validator is different: it checks that content references valid model, Explore and field names.

## System Activity shows use, and datagroups only partly show freshness

Looker admins, and users with the see_system_activity permission, can open the System Activity explores, which are described as connecting to "Looker's underlying application database". Use them to see which Explores and dashboards people open. Google also documents an Unused content folder that helps "quickly identify and manage the unused content" and notifies content owners, which is the only owner reference found.

## Access grants gate Explores, and the ten conditions are laid out below

Per Google's required_access_grants page, it "limit[s] access of an Explore to only those users who have a specific user attribute value assigned to them". That is an access rule an AI agent running as a user inherits, though no per-agent approval flag was found.

Datagroups "synchronize a model's cache retention policy with your database's ETL schedule", so freshness is expressed as a cache policy, not as a freshness test. Each row below names the feature and the Google doc it comes from.

| Condition | What the platform gives you | Gap |
| --- | --- | --- |
| Named owner | Content owners are notified about unused content (Google unused content docs) | No owner field on models or Explores found |
| Documented | description and label in LookML (Google description docs) | Written by developers, not checked for gaps |
| Certified | Content certification and trust levels (Google certification docs) | Manual review by trusted users |
| Quality tests pass | LookML data tests run by Assert Validator (Google CI docs) | Only tests developers wrote |
| Fresh | Datagroups tie cache to ETL schedule (Google caching docs) | A cache policy, not a freshness test |
| Sensitivity classified | none found on 13 Nov 2026 | tags are free text |
| Lineage known | none found on 13 Nov 2026 | Not found in the pages read |
| Active use | System Activity explores; Unused content folder (Google usage docs) | Needs admin or permission |
| Metric defined | LookML measures and Explores (Google LookML docs) | Scope is one model |
| Access policy allows AI use | access_grant and required_access_grants (Google access grant docs) | No approved-for-AI flag found |

## A one-afternoon check in Looker starts from the Explores an agent will query

List the Explores the agent can reach. For each one, record the trust badge, the share of fields with a description, whether data tests exist and passed on the last run, and which access grants apply. Then open the System Activity explores and confirm that real users query the Explore.

Rows that fail go to the model owner as a pull request on the LookML file, where the change is reviewed like any other code.

- Badge: certified or ungoverned.
- Fields: description and label present.
- Tests: Assert Validator run is green.
- Access: required_access_grants set where the data is restricted.

## What Looker leaves to another tool

Column sensitivity labels, upstream lineage and an owner field that agents can read need another tool. Of the 17 tools ranked on this site, the ones listed below come from evidence that names Looker.

Catalog buying and sensitive-data exposure are covered elsewhere.

## 12 of 17 ranked tools name Looker in their evidence

This table is built from the systems each tool names, read 13 Nov 2026. A tool that names Looker can label its assets only as far as its own documentation shows; the marks come from the Readiness Coverage Matrix.

| Rank | Tool | Named as | Reach | Conditions by rule |
| --- | --- | --- | --- | --- |
| 1 | [OvalEdge](/platforms/ovaledge/) | Looker | 5 | 6 of 10 |
| 2 | [Euno](/platforms/euno/) | Looker | 6 | 8 of 10 |
| 3 | [DataHub](/platforms/datahub/) | Looker | 5 | 6 of 10 |
| 4 | [OpenMetadata](/platforms/openmetadata/) | Looker | 6 | 7 of 10 |
| 5 | [Atlan](/platforms/atlan/) | Looker | 6 | 8 of 10 |
| 6 | [Secoda](/platforms/secoda/) | Looker, Looker Studio | 5 | 8 of 10 |
| 9 | [Alation](/platforms/alation/) | Looker | 5 | 7 of 10 |
| 12 | [Monte Carlo](/platforms/monte-carlo/) | Looker | 4 | 5 of 10 |
| 13 | [Snowflake Horizon](/platforms/snowflake/) | Looker | 3 | 6 of 10 |
| 14 | [Precisely](/platforms/precisely/) | Google Looker | 3 | 5 of 10 |
| 15 | [Actian](/platforms/actian/) | Looker | 2 | 6 of 10 |
| 17 | [Collibra](/platforms/collibra/) | Looker | 3 | 5 of 10 |

Tools whose evidence names Looker among the BI tools they read.

## Sources

- Google Cloud: Content certification: https://cloud.google.com/looker/docs/content-certification (read 13 Nov 2026)
- Google Cloud: description (for fields): https://cloud.google.com/looker/docs/reference/param-field-description (read 13 Nov 2026)
- Google Cloud: tags (for fields): https://cloud.google.com/looker/docs/reference/param-field-tags (read 13 Nov 2026)
- Google Cloud: Conversational Analytics best practices: https://cloud.google.com/looker/docs/conversational-analytics-looker-best-practices (read 13 Nov 2026)
- Google Cloud: CI Assert Validator: https://cloud.google.com/looker/docs/ci-assert-validator (read 13 Nov 2026)
- Google Cloud: Content validation: https://cloud.google.com/looker/docs/content-validation (read 13 Nov 2026)
- Google Cloud: Usage reports with System Activity Explores: https://cloud.google.com/looker/docs/usage-reports-with-system-activity-explores (read 13 Nov 2026)
- Google Cloud: Manage unused content: https://cloud.google.com/looker/docs/manage-unused-content (read 13 Nov 2026)
- Google Cloud: required_access_grants: https://cloud.google.com/looker/docs/reference/param-explore-required-access-grants (read 13 Nov 2026)
- Google Cloud: Caching and datagroups: https://cloud.google.com/looker/docs/caching-and-datagroups (read 13 Nov 2026)

Last reviewed 13 Nov 2026. Canonical: https://aireadydatacheck.com/guides/looker/
