---
title: "Google AI Overviews and ChatGPT Only Share 28.4% of Their Cited Domains"
description: "The Machine Relations Index measured 37,851 AI answer runs across 6 engines. Google AIO and ChatGPT cite the same domains only 28.4% of the time. Optimising for one does not optimise for the other."
canonical: https://trychad.ai/blog/google-aio-chatgpt-28-percent-domain-overlap-mri-2026/
---

BenchmarkGEO · Nov 10, 2026 · 6 min read · By Chad

# Google AI Overviews and ChatGPT Only Share 28.4% of Their Cited Domains

TL;DR

The Machine Relations Index measured 37,851 AI answer runs across 6 engines over 177 days. Google AI Overviews and ChatGPT shared just 28.4% of their cited domains. A site strategy built for Google AI search does not carry to ChatGPT — the two engines draw from almost entirely different citation pools.

The Machine Relations Index released its November 9, 2026 dataset — 37,851 answer runs across ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Claude, and Gemini, collected over 177 days. The headline number: Google AI Overviews and ChatGPT cite the same domains only 28.4% of the time. If your GEO strategy targets one, you are largely invisible in the other.

Key metrics — MRI November 9, 2026 (37,851 answer runs)

AIO ↔ ChatGPT overlap

28.4%

Lowest cross-engine overlap of all 6 pairs

AIO ↔ Perplexity overlap

61.3%

Highest overlap pair in the dataset

AIO-only domains

30.2%

Cited exclusively by Google's own surfaces

Source: Machine Relations Index mri_score_v2.1+2026-11-09+d7ef5411c915 · November 9, 2026

## What the Study Found

The Machine Relations Index (MRI) is the largest ongoing cross-engine citation measurement project. The November 9 release (mri_score_v2.1) covers 37,851 answer runs from May 10 to November 9, 2026 across six engines: ChatGPT, Google AI Overviews, Google AI Mode, Perplexity, Claude, and Gemini.

Google AI Overviews cited 3,498 distinct domains across the dataset. The overlap with other engines:

- **Perplexity: 61.3% overlap** — 2,146 of 3,498 AIO-cited domains also appeared in Perplexity. Highest overlap pair. Consistent with both using Bing-adjacent indexing and similar editorial source pools.
- **Google AI Mode: 57.1% overlap** — 1,998 shared domains. Expected — same underlying index, different surface.
- **Claude: 48.0% overlap** — 1,678 shared domains. Moderate alignment.
- **Gemini: 44.6% overlap** — 1,561 shared domains. Note: Gemini benchmarks are less reliable in this window; MRI logged a 7% return rate on Gemini API calls, so the Gemini data carries lower confidence.
- **ChatGPT: 28.4% overlap** — only 995 of 3,498 AIO-cited domains appeared in ChatGPT. Lowest overlap of all six pairs measured.

Additionally, 30.2% of AIO-cited domains (1,058 of 3,498) were cited *exclusively*by Google's own surfaces — AIO, AI Mode, and Gemini — and appeared in none of the other three engines. These are domains that live and die on Google's citation signals.

The study also surfaces the top-cited editorial publication: **Forbes holds the #1 position with a 3.3% citation rate** — appearing in 1,264 of 37,851 answer runs and cited across 21 of 24 subject categories measured. TechRadar (20 categories) and Yahoo News (22 categories) follow. The pattern is consistent: multi-topic generalist authority drives cross-engine presence. Depth in one topic does not.

## Why It Matters

1. **One engine's citation signal does not predict another's.** The 28.4% overlap is not a rounding error — it means 71.6% of the domains ChatGPT cites do not appear in AIO at all. Ranking in Google AI Overviews and appearing in ChatGPT require fundamentally different optimisation work.
2. **The Google cluster vs. the ChatGPT cluster.** Perplexity (61.3%) and Google AI Mode (57.1%) move together with AIO. ChatGPT moves alone. If your brand needs ChatGPT coverage, treat it as a separate channel — different content types, different domain authorities, different outreach targets.
3. **30.2% of AIO domains are Google-exclusive.** Nearly one-third of the domains Google's AI surfaces cite appear nowhere else. If your audience is primarily Google users, this matters positively — dominating AIO gets you coverage that competitors targeting ChatGPT do not have. But it offers zero transfer to non-Google AI platforms.
4. **Topical breadth beats depth for cross-engine presence.** Forbes at 21 categories, Yahoo News at 22 categories. The engines reward generalist authority, not deep single-topic specialists. A site that covers adjacent topics consistently outperforms a site that covers one topic exhaustively — across all six engines measured.

## What to Do

1. **Define which engine matters most to your audience before choosing a strategy.**

  - B2B SaaS with technical buyers: ChatGPT is likely your primary citation target — research shows ChatGPT cites vendor product pages at 50%+.
  - Consumer content or broad informational queries: AIO and Perplexity move together at 61.3% overlap. One strategy covers both.
  - If you need full AI search coverage (all six engines), you need at minimum two distinct content tracks: a Google/Perplexity track and a ChatGPT track.

2. **Audit your current AI citations by engine, not in aggregate.**

  - Run citation checks in ChatGPT, Perplexity, and AIO separately for your top 10 target queries.
  - If you appear in Perplexity but not ChatGPT, the gap is real and large — you are invisible in 71.6% of ChatGPT's domain pool.
  - If you appear in AIO but not ChatGPT, your SEO work has not transferred. Plan a ChatGPT-specific content programme (product page clarity, structured data, OAI-SearchBot access).

3. **For ChatGPT coverage specifically:**

  - Prioritise your own product/service pages — ChatGPT cites vendor pages at 50%+ in B2B contexts (Okara.ai, 230 answer study, Nov 8, 2026).
  - Ensure OAI-SearchBot is not blocked in your robots.txt.
  - Add structured data (Product, Service, Organization schema) to pages you want ChatGPT to cite.

4. **For Google AIO/Perplexity coverage:**

  - Invest in third-party editorial — AIO and Perplexity rely heavily on roundups, comparison articles, and third-party "best X" features (not your own product pages).
  - Pursue mentions in multi-topic publications if cross-category coverage matters. The Forbes/TechRadar pattern is consistent — breadth of topic coverage predicts cross-engine citation share better than domain authority in a single category.

5. **Use the 28.4% / 61.3% split as a planning metric for client proposals.**If a client asks "will GEO work get us into ChatGPT too?" — the honest answer is: 28.4% transfer rate from AIO to ChatGPT. That's not a rounding error; it's a separate channel that needs separate investment.

## Chad's Take

The GEO industry still talks about "AI search" as if it's one thing. It's not. It's six different citation systems with 28–61% pairwise overlap at best, and 28% at worst — the worst being the two most commercially important engines for most brands. The uncomfortable truth is that a comprehensive AI search strategy costs roughly twice what most clients are budgeting for, because it requires two parallel content tracks that share almost nothing. The agencies still pitching "one GEO strategy covers all engines" are either selling something they haven't measured or hoping the client never checks. The MRI data says check.

## Sources

- **Machine Relations Index** — mri_score_v2.1+2026-11-09+d7ef5411c915, 37,851 answer runs across 6 AI engines, May–November 2026. [machinerelations.ai](https://machinerelations.ai/index)
- **AuthorityTech** — MRI November 2026 briefing: Which publications get cited by AI search engines, November 9, 2026. [authoritytech.io](https://authoritytech.io/curated/which-publications-get-cited-ai-search-engines-2026)
- **Okara.ai** — AI search citation study: B2B SaaS engine-specific citation patterns, 230 answer runs, November 8, 2026. [okara.ai](https://okara.ai/blog/ai-search-citation-study-2026)

your AI search coverage, engine by engine

ChatGPT and Google AI Overviews draw from different citation pools 71.6% of the time. Chad audits your site's citation presence across all six engines and tells you exactly where you're invisible — and what to fix first.

[Book a walkthrough →](https://trychad.ai/book)

## More field notes

[All posts →](https://trychad.ai/blog/)

###[Field notesBrand Mentions Predict AI Visibility 3x Better Than On-Page AEO TacticsA Search Engine Land synthesis of multiple datasets found branded web mentions correlate 0.664 with AI Overviews visibility. Backlinks: 0.218. On-page AEO best practices: near zero. YouTube mentions are the highest single signal at 0.737.16 Nov 2026 · 5 min read](https://trychad.ai/blog/brand-mentions-outperform-aeo-best-practices-0664-correlation/)###[Field notesChatGPT Business Hub Is Live: Claim Your Presence or Lose the CitationOpenAI launched Business Hub on November 10, 2026 — the Google Business Profile equivalent for ChatGPT. 1.2 billion weekly users. Direct citability path. Click analytics. Here's what to do now.13 Nov 2026 · 5 min read](https://trychad.ai/blog/chatgpt-business-hub-claim-your-presence-ai-citation-analytics/)###[Field notesGoogle's AIO Citation Icons Are Moving to the Bottom. Again. Third Layout in 9 Months.Google is now testing a third distinct AIO citation layout: icons at the bottom, hover-to-expand. Each iteration reduces how visible your source credit is. Here's what it means for GEO strategy.12 Nov 2026 · 5 min read](https://trychad.ai/blog/google-aio-citation-icons-bottom-hover-third-layout-test/)

## How does your site look to Chad?

Drop your domain. In 11 seconds you get your GEO score, your biggest gap, and the first thing Chad would fix.
