4,040 Google impressions. Zero AI citations.
Our highest-impression blog in Google Search Console earned zero Bing AI citations. Our most-cited AI page earned 444 impressions and 2 clicks. Same site. Same five months. Two scoreboards that do not agree. Zelitho tracks two visibility channels: Google Search Console for classic search appearance, and Bing AI Performance for answer-layer citations. In the audit window, our AI agents in 2025 guide shows 4,040 impressions and 0 AI citations. Our 30-minute GEO checklist (Google ignored it. AI quoted it 3,258 times.) shows 3,258 citations on modest classic search numbers. Optimizing for one channel does not automatically win the other.
Top Google-visible blog
Same URL
GEO audit checklist
Same checklist (2 clicks)
Source: Google Search Console Pages report + Bing AI Performance Pages export (Jun 2026). GSC = classic search impressions and clicks. AI citations = times our URL was visibly shown as a source in AI-generated answers. Different channels, different denominators. Methodology and definitions.
Google winner vs AI winner: same site, opposite scoreboards
The comprehensive AI agents guide is Zelitho’s highest GSC-impression blog in the audit. It is absent from the Bing AI Performance Pages export. The 30-minute GEO checklist is the inverse: modest classic search numbers, dominant AI citation count. Google ignored it. AI quoted it 3,258 times. documents the checklist story. This section puts both URLs on the table. The Google winner has roughly 9.1× more GSC impressions than the AI winner (4,040 ÷ 444). The AI winner has 3,258 citations where the Google winner has 0. Different games.
What each channel actually measures
Two channels. Two winners. Not the same page. Before you compare numbers, understand what each export counts. GSC impressions do not cause or predict AI citations. This is a correlation snapshot from one site in one window, not a controlled experiment.
“We thought our best-performing post in Google was winning everywhere. The AI export showed we were invisible where answers get composed.”
Zelitho content teamGoogle Search Console (classic search)
GSC measures search appearance in Google: how often a URL showed in results, whether users clicked, and average position. High impressions with zero clicks signals visibility without conversion in classic search. Neither URL is a traffic blockbuster in this window, but the AI agents guidedominates impressions among Zelitho blogs.
| Metric | AI-agents guide | GEO checklist |
|---|---|---|
| Impressions | 4,040 | 444 |
| Clicks | 0 | 2 |
| CTR | 0% | 0.45% |
| Avg. position | ~16.6 | ~22.5 |
Bing AI Performance (answer-layer citations)
Bing AI Performance measures citation frequency: times content was visibly referenced in AI-generated answers (Copilot, Bing AI summaries, select partners). Citations are not clicks and not rankings. The AI agents guide does not appear in the Pages export. The checklist is #1 on the site with 54.6% of page-level cites (3,258 of 5,969).
| Metric | AI-agents guide | GEO checklist |
|---|---|---|
| Page-level AI citations | 0 | 3,258 |
| Share of site page-level cites (of 5,969) | 0% | 54.6% |
| Top grounding cluster | n/a | 1,259 |
Why the formats diverge (hypothesis, not proof)
We do not claim the guide failed because it is long-form. We observe that machine-readable structure (checklists, steps, tables) aligns with what Bing’s own guidance recommends and what Google ignored it. AI quoted it 3,258 times. proved on the same checklist URL. The channel gap is real; the format difference is the working hypothesis for remediation.
Format comparison at a glance
Narrative guides can earn impressions when the topic is trending. Checklist assets earn citations when the intent is task-complete. Plot every priority URL on both axes before your next content sprint.
| Dimension | AI-agents guide | GEO checklist |
|---|---|---|
| Structure | Narrative sections, broad coverage | Numbered steps, scannable headers |
| Retrieval fit | Harder to excerpt one block | Easy to quote one audit step |
| Query alignment | Trend topic (“AI agents 2025”) | Task intent (“GEO site audit”) |
| Checklist case study tie-in | Counterexample | Documented winner |
The metric we map content against now
A ranking is a position in a results list. A citation is inclusion in an answer. We were optimizing for one and assuming the other would follow. It did not. Google ignored it. AI quoted it 3,258 times. explained why the checklist won in AI. This case study explains why the team needed a second scoreboard and names the post that won Google but lost AI.
Two scoreboards
GSC tells you search appearance. Bing AI tells you answer-layer inclusion. Both matter for a complete visibility picture.
False confidence
High GSC impressions can mask zero AI citation share on the same URL. The AI agents guide is our clearest example.
Hidden winners
Low-impression pages can carry disproportionate AI citation value. The checklist pattern from Google ignored it. AI quoted it 3,258 times. is proof.
Prioritization
Plot priority URLs on both axes before the next content sprint. Extremes go on the GEO rewrite shortlist first.
“A ranking is a position in a results list. A citation is inclusion in an answer. We were optimizing for one and assuming the other would follow. It did not.”
Zelitho content teamMap every priority URL on both axes
Classic SEO dashboards and AI citation exports answer different questions. Until you merge them on URL, you cannot see which posts look healthy in Google but invisible in Copilot, or which hidden assets deserve more internal links.
Read Google ignored it. AI quoted it 3,258 times. for the checklist anatomy behind the AI winner. This page is the diagnostic frame: the checklist is one column in the side-by-side comparison; the AI agents guide is the inverse.
“We stopped assuming our Google leader was our AI leader. The dual-axis pass made the gap impossible to ignore.”
What changed our content mapHow to run a dual-axis audit (GSC + Bing AI)
Repeatable workflow for teams comparing classic search and AI answer-layer visibility.
- 1Export both channels monthly: GSC Pages (impressions, clicks, position) and Bing AI Performance Pages (citations).
- 2Merge on URL: use path-normalized inventory (
page_inventory.csv) so trailing slashes and domains match. - 3Plot four quadrants: high GSC / high AI (defend); high GSC / low AI (restructure for retrieval); low GSC / high AI (checklist winner pattern, add links); low / low (deprioritize).
- 4Flag extremes: any URL with >1,000 GSC impressions and 0 AI citations goes on the GEO rewrite shortlist.
- 5Borrow the winning format: for high-impression / zero-cite pages, add a checklist or step block at the top without deleting the guide.
- 6Re-export in 30 to 60 days: track citation movement; do not expect instant GSC lift from GEO changes.
Source data
All metrics below come from Google Search Console Pages export and Bing Webmaster Tools AI Performance Pages export, dated June 2026. GSC and Bing AI use different channels and denominators. Do not sum or compare totals across exports as if they measured the same thing.
| Metric | Value | Source |
|---|---|---|
| GSC impressions, AI-agents guide | 4,040 | GSC Pages export |
| GSC clicks, AI-agents guide | 0 | Same |
| GSC avg. position, AI-agents guide | ~16.6 | Same |
| Bing AI citations, AI-agents guide | 0 | Bing AI Performance Pages export (URL absent) |
| GSC impressions, GEO audit checklist | 444 | GSC Pages export |
| GSC clicks, GEO audit checklist | 2 | Same |
| GSC avg. position, GEO audit checklist | ~22.5 | Same |
| Bing AI citations, GEO audit checklist | 3,258 | Bing AI Performance Pages export |
| Impression ratio (GSC winner ÷ AI winner) | ~9.1× | 4,040 ÷ 444 |
| Site page-level AI citation total | 5,969 | Pages export (denominator for 54.6% share) |
| Channels measured | 2 | GSC + Bing AI Performance |
| Date range | Jan 1 to Jun 2, 2026 | Export filenames Jun 2026 |
Quadrant examples from inventory
| Pattern | Example URL | GSC impressions | AI citations |
|---|---|---|---|
| High GSC / zero AI | AI agents in 2025 guide | 4,040 | 0 |
| Low GSC / high AI | GEO audit checklist | 444 | 3,258 |
| High GSC / some AI | AEO metrics post | 1,870 | 463 |
| Mid GSC / minimal AI | Tabbed-content SEO post | 1,102 | 2 |
What this case study does not prove
- Correlation, not causation. We compare two URLs on one site in one window. We do not prove long-form guides cannot earn AI citations.
- Channel scope differs. GSC = Google search. Bing AI Performance = Microsoft AI answer surfaces. Other AI engines (ChatGPT, Perplexity, Gemini) are not in this export.
- Zero clicks ≠ zero value. The AI agents guide may support brand awareness, internal links, or future citations not yet sampled.
- Sampling and aggregation. Bing AI data is sampled; zero in export may mean rare, not impossible.
- Not a format A/B test. We did not republish the AI agents guide as a checklist and remeasure. Remediation steps are hypotheses grounded in Google ignored it. AI quoted it 3,258 times., not experiment results.
Frequently asked questions
Does ranking well in Google mean AI will cite your page?
No. On Zelitho’s site, the highest-GSC-impression blog had 4,040 impressions and 0 Bing AI citations, while a checklist with 444 impressions earned 3,258 citations.
What is the channel gap?
The mismatch between classic search visibility (GSC impressions and clicks) and AI answer-layer visibility (Bing AI citations) on the same URL or across the same content library.
Which data sources did you use?
Google Search Console Pages export and Bing AI Performance Pages export, merged in page_inventory.csv for the audit window Jan 1 to Jun 2, 2026.
Why did the GEO checklist win in AI but not in Google?
Google ignored it. AI quoted it 3,258 times. covers the checklist anatomy. Short answer: scannable audit structure aligned with retrieval intent. GSC and AI measure different surfaces.
Why did the AI-agents guide get Google impressions but zero AI citations?
We observe the split; we do not prove causation. Hypothesis: comprehensive narrative format is harder for AI to excerpt than step-by-step checklists. Republishing with a structured audit block at the top is the remediation test.
How does this relate to “Google ignored it. AI quoted it 3,258 times.”
Google ignored it. AI quoted it 3,258 times. is the AI winner deep dive. This page is the diagnostic frame: that checklist is one column in the side-by-side comparison; the AI agents guide is the inverse.
Can I fix a high-GSC / zero-AI page?
Run the dual-axis audit in the playbook above. Add checklist-style structure, FAQ blocks, and tables without removing useful prose. Re-export Bing AI Performance in 30 to 60 days.
References
Methodology definitions summarized from Microsoft Bing Webmaster Tools AI Performance documentation and Google Search Console help. Anchor IDs match the references section only; body copy links here instead of inline citation markers.
- AI citation and scope. A citation is the total number of times your content was visibly referenced or shown as a source in AI-generated answers during the selected date range. AI Performance reflects activity across Microsoft Copilot, AI-generated summaries in Bing, and select partner AI integrations. It does not measure rankings, authority, performance, or importance.
- Grounding queries. Short grouped phrases representing key terms the AI used when retrieving content that was cited. They are not full user questions or prompts. One grounding query can map to multiple pages; one page can appear under multiple grounding queries.
- Citations vs engagement. A citation indicates that content was visibly referenced in an AI-generated answer. It does not represent traffic, clicks, or user engagement.
- AI vs traditional search metrics. Grounding and citations show how AI systems use your content to generate answers. This differs from traditional search metrics, which measure rankings, clicks, and traffic.
- Content practices for AI inclusion. Align content with user intent; use descriptive headings, concise sections, tables, and FAQ-style structure; strengthen depth across related topics; support claims with evidence; keep content fresh and accurate; maintain consistency across formats.
- Data sampling and views. AI Performance data is aggregated and summarized, not a complete log of every citation. Totals may differ across Pages, Grounding Queries, and timeline views. Timeline trends are observational and cannot be attributed to a single cause or event.
- Google Search Console metrics. Impressions count how often a URL appeared in Google Search results. Clicks count user visits from those results. Average position is the mean ranking position in the export window. GSC measures Google search appearance only; it does not report AI answer-layer citations.
- Measuring on your site. In Bing Webmaster Tools, open AI Performance for your verified site. Export page-level citations as CSV or Excel. In Google Search Console, export the Pages report for the same date range. Merge on normalized URL path for dual-axis analysis.
Definitions above are paraphrased for readability. Export filenames dated June 2026.
Source: Microsoft Bing Webmaster Tools. AI Performance in Bing Webmaster Tools
Your Google leader may be invisible in AI answers.
Run a dual-axis pass on your top 10 URLs, then publish retrieval-ready content that AI can quote. Start with Google ignored it. AI quoted it 3,258 times. or write your first article for $5.