Connecting Search Console to Your Wider AI Visibility Measurement
Combine Search Console generative AI data with GA4, Bing Webmaster Tools and prompt testing to measure visibility in ChatGPT, Claude and more.
Search Console measures Google only. To understand AI visibility in 2026 you need it alongside GA4, Bing Webmaster Tools, prompt testing and server logs.
- Search Console covers Google AI Overviews, AI Mode and Discover AI features.
- ChatGPT, Claude and Perplexity referrals are measured in GA4.
- Bing Webmaster Tools covers Bing and Copilot.
- Mentions need prompt testing.
- Combine sources monthly.
Checked 4 November 2026 (CET) against Search Console Help and Google Search Central documentation. Search Console reports and rollouts change often, and some features only appear for eligible properties, so confirm in your own property. "Our view" marks MentionX opinion. Ledgerline is a fictional company used for illustration.
01How the pieces fit
Use the AI assistant channel in GA4, tracking AI traffic in GA4, AI referral traffic undercounting, checking if ChatGPT, Gemini or Perplexity mention you, measuring AI share of voice, ChatGPT Search brand citations, measuring brand visibility in Claude and Bing Webmaster Tools for AI visibility.
Our view: report Google AI exposure and other assistants side by side.
02The measurement map
| Surface | Where to measure |
|---|---|
| Google AI Overviews and AI Mode | Search Console generative AI report |
| Google Discover AI features | Search Console Discover generative AI report |
| Bing and Copilot | Bing Webmaster Tools |
| ChatGPT, Claude, Perplexity, Gemini app referrals | GA4 with an AI assistant channel |
| Brand mentions in AI answers | Prompt testing and monitoring tools |
| AI crawler activity | Server logs |
03Troubleshooting
| Problem | Cause | Fix |
|---|---|---|
| ChatGPT traffic missing in Search Console | Search Console only covers Google | Use GA4 |
| AI referrals undercounted | Missing referrers | Read about undercounting |
| Mentions without clicks | Answers don't always link | Track mentions separately |
04A monthly AI visibility workflow
- Google AI features. Export the generative AI report by page (lesson 7). Note the top 20 pages and the trend.
- Google organic clicks. Export Performance by page for the same pages. AI feature clicks sit inside these totals.
- AI assistant referrals. In GA4, use the AI assistant channel and landing page dimension. See the AI Assistant channel and custom regex.
- Bing and Copilot. Check Bing Webmaster Tools; see Bing Webmaster Tools for AI visibility.
- Mentions. Run a fixed prompt set in ChatGPT, Claude, Perplexity and Gemini. Record whether you're mentioned, cited or linked, and which competitors appear (as Competitor A-D in reports).
- Crawlers. Check server logs monthly for AI crawler activity; see reading server logs.
- Combine in one sheet with one row per key page and one tab per source.
05Choosing your prompt set
Base prompts on real demand rather than guesses:
- Query groups from Insights (lesson 6) show how people phrase questions in Google.
- Question-style queries from your regex views (lesson 4).
- Comparison queries that name your category, such as "best accounts payable software for mid-size companies".
- Branded questions, such as "is Ledgerline good for multi-entity companies".
Keep the set stable for at least a quarter so trends mean something. See prompt research.
06A combined sheet structure
| Page | Google AI impressions trend | Google organic clicks trend | AI assistant referrals (GA4) | Cited in prompt tests | Notes |
|---|---|---|---|---|---|
| /glossary/three-way-matching/ | Up | Flat | Low | Claude, Perplexity | Definition page |
| /compare/ | Flat | Up | Highest | ChatGPT | Protect |
Use direction labels (up, flat, down) when sharing outside the analytics team.
07Edge cases
- Referrer loss. Some AI apps don't pass referrers, so their traffic lands in Direct. See why AI referral traffic is undercounted.
- Gemini. Gemini app referrals appear in GA4 as a referral source; Gemini isn't part of Search Console's AI features reports.
- Volatile answers. AI assistants vary answers between runs. Run each prompt more than once and record the share of runs that mention you.
- Google-Extended affects some Gemini uses, not Search's AI features; see lesson 8.
08Ledgerline walkthrough
- The SEO lead exported generative AI pages and GA4 AI assistant landing pages for October.
- They ran 20 prompts from query groups in four assistants, three runs each.
- The combined sheet showed the comparison page leading ChatGPT referrals while glossary pages led Google AI impressions.
- The team added a summary table to the comparison page and refreshed glossary entries with clearer definitions.
09Questions practitioners ask
Which number goes to the board? Usually three: non-branded Google clicks, Google AI feature exposure for key pages, and AI assistant referrals with a note on undercounting.
Do mentions without links matter? Yes, for brand demand. Track them alongside referrals; branded search trends (lesson 5) can show their effect.
How often should prompts be run? Monthly is enough for most B2B sites; weekly during a launch.
Should competitors be named in reports? Internally, label them consistently, for example Competitor A-D, so reports can be shared safely.
10The combined sheet, column by column
| Column | Source | Notes |
|---|---|---|
| Month | - | One row per month |
| Non-branded Google clicks | Search Console Performance | Domain property |
| AI feature impressions, key pages | Generative AI performance report | Note any report changes |
| AI assistant referrals | GA4 custom channel group | Undercounts traffic that arrives without a referrer |
| Prompts run | Your prompt library | Same prompts each month |
| Ledgerline mentioned | Prompt results | Count of prompts mentioning the brand |
| Ledgerline cited | Prompt results | Count of prompts linking to your site |
| Competitor A-D mentioned | Prompt results | Labelled, not named |
| Notes | Team | Releases, campaigns, report changes |
11Reading disagreements
| Pattern | Possible reading | Check |
|---|---|---|
| AI exposure up, Google clicks flat | Seen more in AI features, fewer clicks per view | CTR on key pages; content answering the question fully |
| Assistant referrals up, prompt mentions flat | Traffic from a few prompts you don't track | Landing pages for referrals; add prompts |
| Mentions up, referrals flat | Mentioned without links | Add citable pages; track branded search |
| Everything down | Seasonality or a site issue | Indexing, crawl and annotations first |
12Ledgerline walkthrough: a quarterly review
- The SEO lead pulled three months from the sheet. AI feature impressions on the invoice automation guide rose; Google clicks were flat.
- Prompt results showed Ledgerline mentioned in more answers, while Competitor B was cited more often with links.
- The content team added a comparison table and original benchmark method notes to the guide, making it easier to cite.
- The team set a check for the next quarter rather than declaring a result.
Our view: treat the combined sheet as directional. No single source is complete, and the value is in consistent tracking over time.
13Where this fits in the monthly routine
This lesson is the AI section of the monthly summary. Update the sheet in week 1, and write one sentence on what moved and one on what you'll try next.
14Keeping prompts fair
- Use the same wording each month, written the way buyers ask.
- Mix generic category questions with comparison questions.
- Record the assistant, date and whether you were signed in.
- Don't add prompts only after a good result; that skews the trend.
- Retire prompts once a year at most, and note the change.
15Quick checks before you move on
- The combined sheet has at least one complete month.
- Prompts are fixed and recorded with date and assistant.
- Competitors are labelled A-D.
- Stakeholders know each source's blind spots, especially referral undercounting.
16Worked example: Ledgerline's combined view
Ledgerline is a fictional accounts-payable software company. Details and numbers are illustrative.
Ledgerline combined generative AI impressions with GA4 AI referrals. Glossary pages led in Google AI features while comparison pages got most ChatGPT referrals, so the team protected both.
Walkthrough
- Export top pages from the generative AI report.
- In GA4, open your AI assistant channel report.
- Compare the two page lists.
- Run 10 prompts in ChatGPT, Claude and Perplexity and record mentions.
- Record results in your monthly sheet.
17Try it yourself
Build a one-sheet AI visibility view combining generative AI impressions, AI referrals in GA4 and prompt test results for 10 prompts.
18Quick quiz
1. Does Search Console show ChatGPT data?
Answer: No.
2. Where do Bing and Copilot data live?
Answer: Bing Webmaster Tools.
3. How do you see ChatGPT referrals?
Answer: GA4 with a custom AI channel.
4. What shows Google AI Overview exposure?
Answer: The generative AI report.
19What should you check before the next lesson?
- You know where each AI surface is measured.
- You built the combined sheet.
- Read the monthly AI visibility report template.
Next: A monthly Search Console routine. Or return to the Search Console for Marketers track.
20Common questions
Can one tool measure all AI visibility? No. Combine sources.
Is AI referral traffic undercounted? Often, because some apps don't pass referrers.
What about share of voice? Track mentions across prompts.
Questions people ask
3 answersCan one tool measure all AI visibility?
+
No. Combine sources.
Is AI referral traffic undercounted?
+
Often, because some apps don't pass referrers.
What about share of voice?
+
Track mentions across prompts.
Sources
- Search Console Help, Generative AI performance report (Search)support.google.com
- Google Search Central, AI features and your websitedevelopers.google.com
- All accessed 4 Nov 2026.