Recover zero-result retail searches and improve titles

Mine high-intent terms from BigQuery, inspect zero-result queries, and create explainable title recommendations.

Goal

Mine high-intent terms from BigQuery, inspect zero-result queries, and create explainable title recommendations.

Aggregate zero/low-result queries and high-intent terms by category. [source-google-recover-zero-result-retail-searches]
Inspect retail results and competitor title patterns, then draft scored titles. [source-google-recover-zero-result-retail-searches]
A product owner checks truth, brand rules, and prohibited terms before publishing. [source-google-recover-zero-result-retail-searches]

Recompute zero-result rates for the selected queries and verify each title suggestion against actual search terms and product attributes.

If a suggestion merely repeats keywords, inspect query normalization and product mapping; output recommendations only and do not edit live titles.

ADK, BigQuery test catalog/search logs, Playwright test browser, Gemini Computer Use; no live catalog edits.

Reads approved search logs and product data and may write recommendations to an analytics table; do not grant permission to edit live product titles.

Run the source on isolated test data, then check outputs and dependency versions before adapting it.

These values describe the upstream repository, not popularity or likes for an individual skill or playbook.

No metrics snapshot is available yet.