18 / GOOGLE SEARCH CONSOLE

Google Search Console prompts for ChatGPT and AI

Ask for a decision, not an SEO overview. These 18 prompts specify the property, reporting scope and output so your MCP-compatible assistant can turn Search Console data into a useful next step.

ChatGPT

Claude

Notion AI

Google Search Console

Traffic and performance

01Where did clicks fall?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare with the preceding 28 days by page. Sort matched pages by absolute click loss; include clicks and impressions in both periods and percentage changes.

Expected result: A table of the 10 largest observed click losses, with both baselines.
Why use it: Start with losses that affect traffic, rather than a large percentage on two clicks.

Where did clicks fall?

02Which days explain the change?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Query by date and compare with the preceding 28 days. Align weekdays and identify dates with unusually different clicks and impressions.

Expected result: A daily comparison with three dates to investigate and their metric changes.
Why use it: A one-day interruption calls for a different check than a sustained decline.

Which days explain the change?

03Is the decline concentrated in one country?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare country-level clicks and impressions with the preceding 28 days. Rank countries by absolute click change and state their share of current clicks.

Expected result: Five country segments contributing most to the observed change.
Why use it: A local fluctuation should not trigger a rewrite of the whole site.

Pages and queries

04Which pages bring the most visits?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Group by page and sort by clicks. Show impressions, CTR and average position for the top 20 returned pages.

Expected result: A page table with click share relative to the retrieved page sample.
Why use it: Protect valuable pages before chasing tiny opportunities; sample share is not whole-site share.

05Which queries drive this page?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Filter page equals [PAGE URL] and group by query. Sort by clicks; include impressions, CTR and average position.

Expected result: The top 20 returned queries for the exact URL, with metrics.
Why use it: Check what the page already answers before changing its focus.

06Are multiple pages visible for one query?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Filter query equals [QUERY] and group by page. Compare clicks, impressions and average position across returned URLs.

Expected result: A query-to-page table and questions for a manual intent review.
Why use it: Multiple URLs are a clue to investigate, not proof of harmful cannibalization.

SEO opportunities

07Which queries are close to the top 10?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Group by query and page. From the retrieved rows, select average positions 11–20 and at least 100 impressions. Rank by impressions.

Expected result: Ten query–page pairs and the pages to review first.
Why use it: Use an explicit exploratory threshold, not an invented search-volume estimate.

Which queries are close to the top 10?

08Which queries are gaining visibility?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare by query with the preceding 28 days. Rank matched queries by absolute impression growth; include clicks and position in both periods.

Expected result: Ten growing queries, separating visibility growth from click growth.
Why use it: Choose topics with observed demand on your property, not external volume claims.

09What should be refreshed first?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare by page with the preceding 28 days. Identify pages losing clicks while impressions remain stable or rise. Then query their main query–page pairs.

Expected result: Five refresh candidates with evidence and one inspection action each.
Why use it: Check query intent and snippets before assuming the whole page needs rewriting.

CTR and positions

10Where is low CTR worth investigating?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Group by page. From returned rows, select at least 500 impressions and average position 1–10, then sort by CTR. Query each leading candidate by query.

Expected result: Five pages with CTR, impressions, position and the queries to inspect.
Why use it: The 500-impression threshold is a working filter; low CTR at position 40 is a different problem.

Where is low CTR worth investigating?

11Is mobile CTR different?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Group by device for page equals [PAGE URL]. Compare mobile and desktop clicks, impressions, CTR and average position.

Expected result: A two-device table and differences in percentage points for CTR.
Why use it: Check position and query mix before attributing the gap to mobile design.

12Did position improve without more clicks?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare by query with the preceding 28 days. Select matching queries whose average position improved while clicks fell; show impressions and CTR too.

Expected result: Ten conflicting signals with both periods and a hypothesis to check.
Why use it: Average position is not a live rank; a better average alone does not guarantee traffic.

Did position improve without more clicks?

Period comparisons

13What changed between two date ranges?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Override the default dates: compare [START A]–[END A] with [START B]–[END B], using equal lengths and the same web filters, first in total then by page.

Expected result: Clicks, impressions, CTR and position for both periods; absolute and relative click changes.
Why use it: Agree on the baseline before explaining a change; CTR differences use percentage points.

What changed between two date ranges?

14Is the pattern seasonal?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Override the default dates: compare [START]–[END] with [COMPARABLE START LAST YEAR]–[COMPARABLE END LAST YEAR]. Keep country and device filters identical, and explain weekday differences.

Expected result: A year-over-year table with dates and a note on comparability.
Why use it: A repeated pattern can support a seasonal hypothesis; it cannot prove the cause.

Is the pattern seasonal?

15Did a page change coincide with a traffic change?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. For page equals [PAGE URL], compare [BEFORE START]–[BEFORE END] with an equal-length [AFTER START]–[AFTER END] around [CHANGE DATE]. Query by date and query.

Expected result: Before/after metrics and which queries contributed to the change.
Why use it: A before/after comparison documents association, not the causal effect of an edit.

Did a page change coincide with a traffic change?

Reporting and priorities

16Prepare a concise SEO report

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare with the preceding 28 days, in total and by page and query. Write an executive summary, three observed changes and three next actions with supporting figures.

Expected result: A report in the conversation, with dates, scope and evidence for each recommendation.
Why use it: Separate facts from hypotheses; AskGSC does not schedule or email this report.

Prepare a concise SEO report

17Which three actions matter this week?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. Compare with the preceding 28 days by page. Rank candidates using absolute click loss, substantial impressions and average position; request relevant query rows before recommending edits.

Expected result: Three actions with page, evidence, proposed check and reason for priority.
Why use it: Prioritize observed impact; do not invent implementation effort or promised gains.

18What can the available rows actually support?

For [PROPERTY], use Google Search Console web results for the latest complete 28 days. State the exact dates, keep the same filters throughout and distinguish missing rows from zero. List authorized properties, confirm [PROPERTY], then query performance totals and query-level rows for the default period. State retrieved row count and pagination used; do not sum query rows as site totals.

Expected result: A scope note separating totals, returned query sample and missing information.
Why use it: GSC may omit queries; an absent row is not evidence of zero traffic.

What can the available rows actually support?

Connect Search Console

14 days free · No card