Which Engine Cites What: Our Per-Engine Source Data (Sep 30 to Oct 15, 2026)

Name: Which Engine Cites What: Our Per-Engine Source Data
Creator: AEO Spotlight
Published: 2026-10-16
License: https://aeospotlight.com/legal/terms-of-service/

Per-engine source citations from our AI monitoring: 17,223 citations across 7 engines and 5 brands, 30 September to 15 October 2026. Method, limits and SQL on the page.

Published October 16, 2026 · Adam Rock

The finding

AI engines do not share one source diet - but the comparison only holds within one group. Three engines (ChatGPT, Gemini and Perplexity) ran the same 30 questions across 5 brands, and they disagree sharply: ChatGPT and Gemini leaned on Reddit and YouTube, while Perplexity cited neither among its top sources. The other four engines ran a single CRM question, so their source lists are a small sample rather than a profile.

Per-engine source citations - 7 engines (5 brands, 30 Sep - 15 Oct 2026).

Engine	Source citations	Distinct URLs	Brands covered
Perplexity	7,693	1,616	5
Gemini	4,679	1,034	5
ChatGPT	3,917	2,355	5
Grok	382	139	1
Google AI Overviews	215	36	1
Microsoft Copilot	184	20	1
Claude	153	24	1
Total	17,223	-	-

Top sources per engine (top four each):

Engine	#1	#2	#3	#4
ChatGPT	reddit.com (521)	techradar.com (420)	en.wikipedia.org (113)	youtube.com (92)
Gemini	youtube.com (498)	reddit.com (383)	technologyadvice.com (105)	runrepeat.com (103)
Google AI Overviews	youtube.com (56)	reddit.com (31)	forbes.com (20)	-
Perplexity	runnersworld.com (324)	forbes.com (256)	clickup.com (236)	rtings.com (211)
Microsoft Copilot	forbes.com (17)	fitsmallbusiness.com (17)	crm.org (17)	capterra.com (16)
Grok	itechguides.com (19)	zdnet.com (17)	pcmag.com (15)	forbes.com (13)
Claude	techradar.com (33)	stackshare.io (20)	fayedigital.com (17)	fitsmallbusiness.com (10)

Universal vs engine-specific sources

Some publishers are cited by nearly every engine; others belong to one. forbes.com was the only source cited by all seven engines. zapier.com and expertmarket.com appeared on six. Twelve sources appeared on five, including reddit.com, techradar.com, monday.com, hubspot.com, zoho.com and salesforce.com; youtube.com reached four.

At the other end, en.wikipedia.org drew 113 citations on exactly one engine (ChatGPT), and the specialist review and benchmark sites are just as narrow: runnersworld.com on two engines, rtings.com and runrepeat.com on three.

The record-type split

The headline figure is 17,223 source citations, not 18,943 total records. The source table counts only citations to a publication; the rest of the ledger is brand mentions with no cited URL and organic results.

Record type	Count
Source citations (info_site)	17,223
Brand mentions with no cited URL (ai_response)	1,570
Organic results (organic_search)	150
Total records	18,943

Sentiment

Of the records we scored, 17,683 were neutral, 1,242 positive and 18 negative - roughly 93% neutral. Read that carefully: the classifier is keyword-based with a deliberately wide neutral band, so neutral is the default bucket, not a precision measure. The defensible reading is that AI rarely says anything strongly polarised about a brand; the failure mode we actually see is absence from the answer, not negative sentiment.

Method

Universe: 5 brands, one per category (CRM, project management, ecommerce, AI-visibility tools, running shoes), 6 buying questions each - 30 questions - run daily over 16 days (30 Sep - 15 Oct 2026).
Engines: ChatGPT, Gemini, Google AI Overviews, Perplexity, Claude, Grok and Microsoft Copilot - the seven engines that produced source citations in the window. Three of them (ChatGPT, Gemini, Perplexity) ran all 30 questions; the other four ran a single CRM question.
Window: 30 September 2026 to 15 October 2026 (local time).
Measure: citations the engines made to a publication (classified as info_site), counted per engine.
Sample: 17,223 source citations out of 18,943 total records.
Roster change mid-window: Brave AI left the monitored set on 2026-10-04 (near-zero citation yield - it mostly cited its own domain) and DeepSeek took the slot.

Limitations

Uneven coverage. Three engines (ChatGPT, Gemini, Perplexity) ran all 30 questions across all 5 brands, so those three are directly comparable to each other. The other four engines ran one CRM question, so their rows are a small sample and are not comparable to the first three.
Rows, not pages. Counts are raw citation rows. Distinct URLs are reported separately, which is why ChatGPT shows 2,355 distinct URLs against 20 for Microsoft Copilot.
Sentiment is coarse. A keyword model with a wide neutral band; neutral is the default bucket. Do not read the ~93% as a precision measure.
Claude was under-extracted. An extraction bug dropped the Claude sources for the whole window; 153 rows were recovered from the stored raw responses. Treat single-engine numbers as provisional.
Engine set. This window covers the 7 engines that ran. DeepSeek joins the set next week and its cut will be published once it has real data behind it.
No paid placements were involved, and we make no claim that any of this causes citations.

How we computed it

Be clear about what this is: the query below reads our internal measurement database - the research tenant behind this data - which readers do not have access to. It is published as the audit trail for how the numbers on this page were produced, not as something you can execute yourself. A method you cannot inspect is just an assertion, so it is here to be inspected.

To check the arithmetic, the per-engine figures and the method box are in the machine-readable file at /data/ai-citation-sources-per-engine-2026-10/dataset.json. Check it yourself: the same file ships sourcesAll - all 1,489 source hosts, each with how many of the seven engines cited it, plus the citations and distinct URLs per engine and the total. Sum the citations column for 17,223; sum each engine's column for its per-engine total (Perplexity 7,693, Gemini 4,679, ChatGPT 3,917, Grok 382, Google AI Overviews 215, Microsoft Copilot 184, Claude 153); count rows with engines = 7 for the single universal source and engines = 1 for the 1,163 that exactly one engine cited. Checking the arithmetic without taking our word for it is the point of publishing it.

Two notes if you build something similar. The Date column is a bare datetime(6) holding UTC, so DATE(Date) is a UTC day - though some MySQL clients render it shifted by the session time zone, which is a display artifact rather than a storage convention. And the per-engine counts are rows in an answers table, not deduped by page, so a count of 7,693 is genuinely 7,693 citation rows.

SELECT Platform,
                       COUNT(*)                  AS SourceCitations,
                       COUNT(DISTINCT Url)       AS DistinctUrls,
                       COUNT(DISTINCT BrandName) AS BrandsCovered
                FROM   tracker_citations
                WHERE  TenantId = 9001
                  AND  CitationSourceType = 'info_site'
                  AND  Platform <> 'Brave AI'
                  AND  Date >= '2026-09-30' AND Date < '2026-10-16'
                GROUP  BY Platform
                ORDER  BY SourceCitations DESC;

Top sources per engine (top four each):

SELECT Platform, Host, c, rn FROM (
                  SELECT Platform,
                         REPLACE(SUBSTRING_INDEX(SUBSTRING_INDEX(Url, '/', 3), '/', -1), 'www.', '') AS Host,
                         COUNT(*) AS c,
                         ROW_NUMBER() OVER (PARTITION BY Platform ORDER BY COUNT(*) DESC) AS rn
                  FROM   tracker_citations
                  WHERE  TenantId = 9001 AND CitationSourceType = 'info_site' AND Platform <> 'Brave AI'
                    AND  Date >= '2026-09-30' AND Date < '2026-10-16'
                  GROUP  BY Platform, Host
                ) t
                WHERE rn <= 4
                ORDER BY Platform, c DESC;

Record-type split and sentiment:

SELECT CitationSourceType, COUNT(*) AS c FROM tracker_citations
                WHERE TenantId = 9001 AND Platform <> 'Brave AI'
                  AND Date >= '2026-09-30' AND Date < '2026-10-16'
                GROUP BY CitationSourceType;

                SELECT Sentiment, COUNT(*) AS c FROM tracker_citations
                WHERE TenantId = 9001 AND Platform <> 'Brave AI'
                  AND Date >= '2026-09-30' AND Date < '2026-10-16'
                GROUP BY Sentiment;

The machine-readable version of this dataset is at /data/ai-citation-sources-per-engine-2026-10/dataset.json.

Read the narrative version of this finding: Which engine cites what - the analysis.

Cite this

AEO Spotlight, Which Engine Cites What: Our Per-Engine Source Data (30 September to 15 October 2026). Retrieved from https://aeospotlight.com/data/ai-citation-sources-per-engine-2026-10

Frequently Asked Questions

What is this dataset?

A per-engine count of the sources cited in AI answers from 30 September to 15 October 2026: 17,223 source citations out of 18,943 records, across 7 engines and 5 brands.

Which engines were measured?

Seven: ChatGPT, Gemini, Google AI Overviews, Perplexity, Claude, Grok and Microsoft Copilot. Brave AI was excluded because it only ever cited its own domain.

Why is the headline 17,223 and not 18,943?

18,943 is every citation record. 17,223 of those are source citations (info_site); the other 1,720 are 1,570 brand mentions with no cited URL and 150 organic results.

Can I cite this data?

Yes. Please credit AEO Spotlight and link to this page. A suggested citation is at the bottom of the page.