AI Search SERP Feature Citation Map: Where AI Mentions Appear in 2026

TL;DR

AI citations no longer live in a single "sources" box. By 2026 they span at least nine distinct surfaces — inline footnotes, source carousels, follow-up panels, related questions, in-product action chips, and voice answer attributions — each with its own citation format. Tracking only AI Overviews misses 50-70% of where your brand actually appears.

How to use this checklist

Work through each surface below in order. For every item:

1. Run the detection query in the listed engine.
2. Confirm whether your domain is currently cited.
3. Log the surface, the prompt, and the cited URL into your GEO tracking sheet.
4. Tag any missing surface as a citation gap and assign it to a content owner.

A full audit usually surfaces 4-6 untapped surfaces per topic cluster.

The 9 AI citation surfaces (2026)

1. Google AI Overviews — inline citation chips

• Where it appears: Above the organic 10 blue links on supported informational queries.
• Citation format: Small clickable chips with favicon + domain, plus an expandable "Show more" tray that reveals 6-13 sources.
• Detection: Search the target query in Google with udm= parameter cleared. If an AI Overview renders, expand the source tray.
• Audit signal: Position within the tray (top 3 vs. "Show more" hidden) is the dominant CTR driver.
• Common pitfall: Many tracking tools only capture the visible 3 chips; always expand the tray.

2. Google AI Mode — conversational source cards

• Where it appears: Inside the dedicated AI Mode tab (google.com/ai) reached via the AI Mode toggle on mobile.
• Citation format: Vertical source cards stacked beside the answer, each with publisher logo, headline, and snippet.
• Detection: Open AI Mode, ask a follow-up question, then scroll the right rail or below-the-fold card stack.
• Audit signal: AI Mode pulls from a wider domain set than AI Overviews — Reddit, YouTube, and forum threads frequently surface here even when absent from Overviews.

3. Perplexity — numbered inline citations + Sources tab

• Where it appears: Inside every Perplexity answer (web app, mobile app, Comet browser).
• Citation format: Bracketed numbers [1] [2] [3] inline, plus a Sources panel listing 4-8 cited URLs with publish date and short excerpt.
• Detection: Run query, click the Sources tab, then click "View all" if available.
• Audit signal: Perplexity cites 3-6 domains per answer on average and leans heavily on news, academic, and primary docs.

4. Perplexity — Related & follow-up surfaces

• Where it appears: Below the main answer (Related questions) and inside Discover/Spaces.
• Citation format: Each related answer reuses the numbered citation pattern, generating additional citation slots per topic.
• Detection: Click 2-3 related questions for the target topic and inspect their Sources tabs.
• Audit signal: Related questions often draw from a different domain set than the parent query — track them separately.

5. ChatGPT Search — inline link annotations

• Where it appears: Inside ChatGPT (web, desktop, mobile) when the model performs a web search.
• Citation format: Hyperlinked phrases inside the answer text plus a collapsible "Sources" footer listing 4-10 URLs.
• Detection: Prompt with explicit "search the web for…" or use the search-enabled icon, then expand the Sources footer.
• Audit signal: ChatGPT skews toward reference sites, official docs, and Reddit; brand mentions in inline annotations carry more weight than footer-only citations.

6. Microsoft Copilot — numbered footnote citations

• Where it appears: Copilot in Bing, Windows Copilot, Microsoft 365 Copilot Web, and Edge sidebar.
• Citation format: Numbered superscripts ¹ ² ³ inline, with a footer revealing favicon, title, and URL.
• Detection: Run query in Bing Copilot or copilot.microsoft.com, hover each superscript to confirm domain.
• Audit signal: Copilot frequently cites Bing-indexed pages that don't appear in Google's AI Overviews — index parity is the prerequisite.

7. Gemini — citation links + "Double-check" verification

• Where it appears: Inside the Gemini app and Gemini-powered Google Workspace surfaces (Docs, Gmail).
• Citation format: Underlined inline links plus an optional "Double-check response" pass that highlights green-supported and orange-conflicting passages with source links.
• Detection: Click "Double-check response" in Gemini for any factual answer.
• Audit signal: Double-check often reveals citation slots not shown by default — track both default citations and verification citations.

8. Claude — Projects & web-search source cards

• Where it appears: Inside Claude (claude.ai) when web search or a Project knowledge base is active.
• Citation format: Source cards rendered below the response with title, domain, snippet, and the prompt that triggered retrieval.
• Detection: Enable web search in Claude, run query, expand the Sources panel.
• Audit signal: Claude favors longer-form authoritative sources and frequently de-duplicates against the same domain — diversifying URLs per concept improves coverage.

9. Voice & smart-speaker spoken answers

• Where it appears: Google Assistant, Alexa, Siri, ChatGPT Voice, Perplexity Voice.
• Citation format: Spoken attribution ("According to …") with the source name; visual citation appears on companion screens.
• Detection: Speak the target query into each assistant; record the spoken source name and the screen card.
• Audit signal: Voice attribution is winner-take-one — only the top-ranked source is read aloud, making this the highest-stakes surface per query.

Cross-surface tracking checklist

For every priority topic, confirm you have:

• [ ] Captured AI Overviews expanded source tray (not just visible chips).
• [ ] Logged AI Mode card stack on mobile and desktop.
• [ ] Pulled Perplexity Sources tab for both the main and 2 related questions.
• [ ] Recorded ChatGPT Search inline + footer citations.
• [ ] Inspected Copilot answer in Bing and Edge sidebar.
• [ ] Run Gemini Double-check on factual queries.
• [ ] Tested Claude with web search enabled.
• [ ] Played the query through at least one voice assistant.
• [ ] Tagged each missing surface in your quarterly GEO audit.

Misconceptions

• "AI Overviews = AI search." AI Overviews are one of nine surfaces, and often the least diverse. Most enterprise GEO programs underperform because they only track Overviews.
• "Citations and rankings move together." Multiple 2026 studies show AI Overview citations drifting away from page-one organic rankings, with many cited URLs ranking outside the top 10.
• "More citations always means more traffic." Many surfaces are zero-click by design. Track citation share and assisted conversions, not just sessions.

How to apply

1. Inventory once per quarter. Walk this checklist across your top 25 priority queries.
2. Map gaps to content fixes. A missing Perplexity citation usually means thin entity coverage; a missing Copilot citation often means a Bing indexing gap.
3. Instrument tracking. Use AI citation tracking tools or scripted scrapes to monitor each surface weekly.
4. Report citation share, not citation count. Share is comparable across queries; count is not.

FAQ

Q: How many AI citation surfaces should I track in 2026?

At minimum nine: Google AI Overviews, Google AI Mode, Perplexity main answer, Perplexity related questions, ChatGPT Search, Microsoft Copilot, Gemini (default + Double-check), Claude with web search, and at least one voice assistant. Programs that track fewer than five typically miss the majority of brand citations.

Q: Which surface drives the most referral traffic?

Perplexity and ChatGPT Search currently produce the highest click-through rate per citation, while AI Overviews drive the largest impression volume but the lowest CTR per citation. Voice surfaces drive almost no clicks but high brand recall.

Q: Can I use a single tool to track every surface?

No single tool covers all nine surfaces well in 2026. Most enterprise GEO stacks combine an AI Overviews tracker (e.g., Semrush, Ahrefs, ZipTie), a multi-engine prompt monitor (e.g., Profound, Peec, Otterly), and manual sampling for voice assistants.

Q: How often do AI citation surfaces change?

Formats and source-tray sizes change roughly every 6-8 weeks. Re-validate your detection patterns at least quarterly and after every public engine update.

Q: Is being cited the same as being linked?

Not always. Some surfaces (voice answers, Gemini Double-check) attribute by source name without a clickable link. Track both linked and unlinked attributions to capture full brand exposure.