How do AI search engines behave for your category and market? Consult our GEO Compass dataset.

This data is sourced from our deep experience running AI search programs. If you have any questions or want to see how your brand appears, contact AIAdvisors@wecommunications.com.

1.2M
AI answers analysed
8.0M
source citations classified
73k
brands and organisations named
11
AI engines and models
Aug 25 – Nov 26
24 markets · 21 languages
Try it

What does AI cite in your category and market?

Pick a market on the map, then an industry, AI engine or question type. The results update as you choose.

Pick a market

Showing all markets. Click a market or region to filter the results.

Six things every brand should know about AI search

Trend
99.5%of ChatGPT answers used live web search in Aug 2026

Almost all AI answers now consult live web data

On a fixed panel of long-running studies, the share of ChatGPT answers that ran a live web search rose from 35% in September 2025 to 99.5% in August 2026. Gemini went from 15% to 83%. What sits on the open web, and who it credits, now shapes almost every answer.

Share of answers that ran a live web search before answering, by engine. Same set of long-running studies every month.
Sources
32%of citations come from brand and corporate websites

Brand websites are cited more than earned, social, and institutional sources

Company and brand websites supply 32% of the distinct domains AI engines cite, and 82% of answers with sources cite at least one. The other 68% is earned media (21%), academic and institutional sources (15%), social platforms (8%) and a long tail of reference, government and marketplace sites.

Branded vs. Unbranded
70% vs 13%of answers cite the brand’s own site: branded vs unbranded prompts

When brands are named in prompts, their website is cited in 70% of answers. For unbranded prompts, this drops to 13%.

In unbranded questions the focus brand is named in 28% of answers, but its own website is cited in only 13% of the answers that list sources. When the question names the brand, own-site citation jumps to 70%. Most category discovery happens on other people’s pages.

Trend
27.6% → 6.5%answers retrieving Reddit, Sep 2025 → Oct 2026

Reddit’s share of the answer is shrinking

Across a fixed panel of studies tracked since September 2025, the share of answers retrieving Reddit fell from 27.6% to 6.5% by October 2026, and YouTube from 28.7% to 13.1%. LinkedIn climbed from almost zero to 3.2% in August before slipping back to 1.9%. Community content still matters, but engines are leaning on it less every quarter.

Share of answers retrieving each platform. Same set of long-running studies every month.
Engines
39%of Google AI Overviews answers cite YouTube

Google AI Overviews runs on YouTube

YouTube appears in 39% of Google AI Overviews answers with sources, and in 40–47% of them each month since July. Google AI Mode leaned on it almost as much until it cut its sources in late August 2026, falling from 45% to 16% by October. On ChatGPT it is 3%, on Gemini 4% and on Claude about 1%.

Trend
71% → 58%of ChatGPT answers carried shopping modules, Aug → Oct 2026

Shopping modules surged, then cooled

On a fixed panel, ChatGPT attached shopping or product modules to 71% of answers in August 2026, up from 32% a year earlier, before easing to 58% in October. Google AI Mode went from none to 44% in July and AI Overviews to 35% by September, and both settled near 31% in October. ChatGPT showed ads in about 12% of answers in July and August 2026; none have been detected on the panel since September.

Share of answers showing shopping or product cards, by engine. Same set of long-running studies every month.
Branded vs. Unbranded

Being named when someone asks for you is easy. Being recommended when they don’t is the real test.

70%
of branded answers cite the brand’s own site

Branded questions narrow the source set

Answers to questions that name a brand cite 6.2 distinct domains on average, against 8.3 for unbranded questions. The brand’s own website is cited in 70% of those answers, which makes owned pages the default fact base once a buyer knows your name.

24%
of branded-answer citations are earned media

Reputation is decided off-site

Even in branded answers, earned media makes up 24% of citations (20% for unbranded) and social and community platforms 10% (7%). Wikipedia appears in 8% of branded answers against 5% of unbranded. What others publish about a brand travels into its own branded answers.

10.3%
of brand-trust answers frame brands negatively

Negative framing concentrates in trust and health questions

Across the dataset, 5% of answers that discuss brands frame them negatively. That rises to 10.3% for brand trust and reputation questions and 15.7% for health and medical questions, against 2.6% for recommendation questions.

AI engines

Every engine builds its answers differently

Engines
14 → 4
sources per Google AI Mode answer, before and after late August 2026

Google AI Mode cut its sources by two-thirds

In the last week of August 2026, Google AI Mode went from about 14 distinct domains per answer to about 4, and it has stayed there into November. It shows up in the same week across studies tracked on two independent platforms, so it reflects a change in Google’s product rather than in measurement. In October AI Mode averaged 3.7 sources per answer, against 7.7 for Perplexity, 7.5 for ChatGPT and 6.4 for Google AI Overviews; only Gemini (3.2) cites fewer. Each remaining AI Mode citation now carries far more weight.

Engines
22%
of OpenAI API answers cite Wikipedia

Each engine has its favourite platforms

YouTube leads on Google AI Overviews, and LinkedIn shows up in 13% of Perplexity answers and 12% of AI Mode answers. The OpenAI API cites Wikipedia in 22% of answers, against 7% for the ChatGPT app. Claude almost never cites Reddit or YouTube.

Engines
42% → 30%
earned-media share of Claude citations, Sonnet 4.6 → 5

Model upgrades move citations overnight

Within the same studies, the switch from Claude Sonnet 4.6 to Sonnet 5 in August 2026 cut earned media’s share of citations from 42% to 30%, and sources per answer dipped from 6.6 to 6.2. Visibility has to be tracked by model version, not just by engine.

Insights

Latest analysis

AI search statistics 2026: usage, clicks, citations and GEO, with sources

The key AI search statistics for 2026: ChatGPT, Gemini and AI Overviews usage, the click impact of AI summaries, what AI engines cite, and country data for Australia, Singapore, Malaysia, Indonesia, India and Europe. Every figure linked to its source.

What is GEO? Generative engine optimisation explained, with data

Generative engine optimisation (GEO) is how brands earn a place in AI answers from ChatGPT, Gemini, Perplexity and Google AI Overviews. What it is, how it differs from SEO, and what 900,000+ AI answers show about how it works.

The GEO checklist: 25 ways to improve your brand's AI visibility

A practical generative engine optimisation checklist covering your website, earned media, social and video, reference sites and measurement, grounded in what the GEO Compass benchmark shows AI engines actually cite.

AI search in Asia-Pacific: how GEO differs in Australia, Singapore, Malaysia, Indonesia and India

AI engines cite different sources depending on where a question is asked. A market-by-market comparison of AI search in Australia, Singapore, Malaysia, Indonesia and India, from the GEO Compass benchmark.

ChatGPT vs Gemini vs Perplexity vs Google AI Overviews: how each AI engine cites sources

How many sources each AI engine cites, which platforms it favours and how often it recommends brands, from 900,000+ answers across ChatGPT, Gemini, Perplexity, Claude, Copilot and Google AI Overviews and AI Mode.

How to measure AI visibility: share of model and the new GEO KPIs

Clicks and impressions don't capture how AI engines describe your brand. The KPIs for AI visibility: share of model, answer position, sentiment, accuracy, citation share and AI referrals, with benchmarks from 900,000+ AI answers.

Your brand in AI search

How is your brand showing up in AI answers?

We benchmark how ChatGPT, Gemini, Google AI Overviews, Perplexity, Claude and Copilot describe, cite and recommend brands, then build the content and earned-media plan to change it.

Want to understand your brand opportunities and risks on AI search? Contact We AI & Data Advisors for a sample audit: AIAdvisors@wecommunications.com