Find High Intent Queries With Google Search Console AI

Google Search Console AI-Powered Configuration — What It Is

Official announcement: Google Search Central Blog, December 4, 2025 — https://developers.google.com/search/blog/2025/12/ai-powered-configuration

Full global rollout confirmed: February 18–19, 2026 (Search Engine Land, Mediology Software, PikaSEO)

Current status (March 2026): Live for all verified Search Console properties worldwide. If you don't see it, hard-refresh your Performance report or check a different property.

What it is: A natural-language prompt interface layered on top of the existing Search results Performance report. It is NOT a new AI analytics engine — it is a configuration shortcut. The AI reads your prompt and applies the appropriate filters, date comparisons, and metric toggles automatically. The underlying data (clicks, impressions, CTR, average position) is identical to what was always there.

Daily limit: 20 AI prompt requests per day per property.

Cost: Free, included in all Google Search Console accounts (no paid tier required).

Official help doc: https://developers.google.com/search/blog/2025/12/ai-powered-configuration

How It Works — Technical Detail

The feature handles exactly three tasks automatically:
1. Metric selection — toggles Clicks, Impressions, Average CTR, and Average Position based on your prompt
2. Filter application — narrows data by query, page, country, device, search appearance, or date range
3. Comparison configuration — sets up date-range comparisons (e.g., Q1 2026 vs Q1 2025) without manual date entry

What it cannot do:
- Sort tables
- Export data to CSV
- Perform regex filtering (confirmed by early tester Pedro Dias on LinkedIn: "it applies basic filtering most times")
- Save report configurations across sessions — each prompt generates a temporary view that resets on navigation
- Submit URLs for indexing, change permissions, or perform any administrative action
- Support Discover or Google News performance data
- Handle complex multi-dimensional layered queries reliably (e.g., filtering by region AND specific metric simultaneously)

How the 'bottom of funnel queries' prompt actually works: The AI interprets your prompt and applies a query-contains filter for transactional/commercial intent keywords (buy, price, deal, discount, coupon, review, best, versus, comparison, download, sign up, get a quote, contact, near me, shipping, cost, order, purchase, subscribe, trial, demo). It does NOT use machine learning to classify intent — it applies keyword-match filters. Always verify the filters it applied before trusting the output.

Architecture note: Google has not disclosed which underlying LLM powers the configuration feature. It is part of Google's broader push to embed Gemini-class AI across its webmaster tools suite.

Try It Yourself

Prerequisites

• A verified Google Search Console property with at least 28 days of data
• A Google account with Owner or Full User permissions on the property
• No install required — browser-based at https://search.google.com/search-console

Step-by-Step Walkthrough

Step 1 — Access the feature

1. Go to https://search.google.com/search-console
2. Select your property from the top-left dropdown
3. Click "Performance" in the left sidebar
4. Click "Search results" (default tab)
5. Look for the banner: "New! Customize your Performance report using AI."
   OR click the filter/sparkle icon in the report header to open the AI side panel

Step 2 — Run the bottom-of-funnel prompt

Type into the AI prompt field:
"Show me bottom of funnel queries"

OR for more precision:
"Show me queries containing words like buy, price, review, discount, 
versus, or near me in the last 3 months"

Step 3 — Verify the filters applied

Before analyzing: check the filter bar at the top of the report.
Confirm it shows something like:
  Query > Contains > [buy OR price OR review OR...]
If the filter looks wrong, clear it and rephrase your prompt.

Step 4 — Add position data and filter

1. Click "Average position" toggle at the top of the report to enable it
2. Click "+ New" in the filter bar
3. Select "Position" > "Greater than" > type "3"
4. Apply the filter
5. Optional: also filter Position < 20 to focus on pages 1-2 of Google

Step 5 — Sort and prioritize

1. Click the "Position" column header to sort ascending (best positions first)
   OR sort by "Impressions" descending to find highest-volume opportunities
2. Change "Rows per page" (bottom of table) to 500
3. Look for: High impressions + Position 4–20 + Low CTR = highest-priority targets

Step 6 — Click into a query to find the ranking page

1. Click any query in the table
2. Click the "Pages" tab in the detail view
3. This shows which URL on your site is ranking for that query
4. Open that page and add the query term prominently (H2, intro paragraph, or FAQ section)

Step 7 — Export for tracking (manual step — AI cannot do this)

1. Click the Export button (top-right of the Performance report)
2. Choose Google Sheets or CSV
3. Track these queries monthly to measure position improvement

Additional High-Value Prompts to Try

"Show me queries with high impressions but low CTR in the last 28 days"
"Compare mobile vs desktop performance for the last 30 days"
"Show me question-based queries containing 'how to' or 'what is'"
"Show me queries containing 'alternative' or 'vs' or 'comparison'"
"Compare Q1 2026 to Q1 2025 for all queries"
"Show me queries from mobile devices in the United States"

Prompt Best Practices

✅ Be specific: "last 28 days" not "recently"
✅ Use GSC terminology: "average position" not "ranking"
✅ Include all dimensions in one prompt: "mobile clicks from Germany for /blog pages"
✅ Always verify the applied filters before analyzing
❌ Don't expect regex — it applies basic keyword matching only
❌ Don't use it for Discover or News data — not supported
❌ Don't exceed 20 prompts/day or you'll hit the daily limit

What The Creator Didn't Mention

1. The AI is a filter configurator, not an analyst

The feature does NOT analyze your data with AI intelligence. It translates your prompt into standard GSC filters. The insight comes from you interpreting the filtered data — not from the AI. Google explicitly warns that the AI can misinterpret requests and recommends always reviewing the applied filters.

2. The 20-request daily cap

Users are limited to 20 AI prompt requests per day. High-volume agencies managing multiple client accounts may find this restrictive during audits or month-end reporting.

3. No saved configurations

Each AI-generated report view resets when you navigate away. There is no way to save a prompt as a recurring report. You must re-enter prompts each session.

4. The 'just use the keywords on your page' claim is oversimplified

The creator implies that placing high-intent keywords prominently on a page is sufficient to move from position 4–20 to top 3. In reality, ranking improvement depends on: page authority, backlink profile, content depth, E-E-A-T signals, technical SEO, and competition. Keyword placement is necessary but rarely sufficient alone.

5. AI Overviews are crushing CTR for informational queries — but BOFU queries are more resilient

Ahrefs (December 2025, 300K keywords) found AI Overviews reduce position-1 CTR by 58%. However, this impact is concentrated on informational queries. Transactional/commercial queries (exactly what BOFU targeting captures) are less affected by AI Overviews because Google tends to show shopping results, ads, and direct links for purchase-intent searches. This actually makes the creator's BOFU strategy more valuable in the current environment, not less.

6. Alternatives to this GSC AI feature for BOFU keyword discovery

• Semrush / Ahrefs — filter keyword database by intent type (transactional/commercial) with volume and difficulty data the GSC AI cannot provide
• Google Keyword Planner — free, shows search volume for transactional terms
• Manual GSC regex filter — in the standard filter bar, use regex: buy|price|review|discount|vs|alternative|coupon|near me — this achieves the same result without using your 20 daily AI prompts
• Screaming Frog + GSC API — for bulk analysis across large sites

7. The feature does not work for Discover or Google News

Publishers who rely heavily on Discover traffic cannot use this feature for those report types. It is Search results only.

8. compactkeywords.com is a paid course

The creator promotes their own paid SEO course at the end of the video. The GSC AI feature itself is completely free and requires no third-party tools or courses to use.