What is the GEO Brand Citation Index?

The GEO Brand Citation Index is a monthly dataset that measures how frequently brands are cited by ChatGPT, Perplexity, and Gemini when asked common evaluation questions about tools and services. Each brand receives a normalised score from 0 to 100 per platform. The gap between platforms — the delta — reveals whether a brand is cited from AI training data or from the current live web.

The index was built by Artur Ferreira at The GEO Lab because brand teams had no way to measure their AI citation visibility — only to guess at it. Traditional SEO tools track keyword rankings in Google. They tell you nothing about whether ChatGPT recommends you when someone asks “what CRM should I use?” or whether Perplexity surfaces you when someone asks “what are the best AI writing tools?”

The index fills that gap. It is publicly available, methodologically transparent, updated monthly with the same fixed query panel, and designed so that the numbers are directly comparable over time. Full methodology including scoring formula, normalisation, and academic references →

Related on The GEO Lab

What is an AI Memory Brand?

An AI Memory Brand is a brand that scores significantly higher on ChatGPT than on Perplexity. It is well cited in AI training data — recommended by language models because it was prominent in industry content before the model’s knowledge cutoff — but its current live web presence is weaker than its AI reputation suggests.
🧠 Definition: AI Memory Brand

An AI Memory Brand is a brand whose AI citation visibility is based on historical training data rather than current web presence. ChatGPT is trained on a fixed dataset. When it recommends a brand, it does so because that brand appeared frequently in SEO guides, review roundups, and industry content before its training cutoff — which may be one to two years ago. Perplexity retrieves the live web in real time. When a brand scores high on ChatGPT but significantly lower on Perplexity, the specific conclusion is: the brand exists in AI memory, but is not being actively written about, linked to, or cited at the same level on the live web today.

The risk is structural. AI training data updates over time. A brand that fails to maintain live web presence will eventually lose its training data advantage too — because the content that taught ChatGPT to recommend it will be outnumbered by newer content that doesn’t.

In March 2026, the clearest AI Memory Brands were Ahrefs (ChatGPT: 100.0, Perplexity: 51.9, delta: −48.1) and Google Search Console (ChatGPT: 61.9, Perplexity: 33.3, delta: −28.6). Both were standard references in SEO content for years. ChatGPT learned from that historical corpus. Perplexity reads the web as it exists in 2026, and the gap is measurable and timestamped.

All 28 brand scores — March 2026

Complete normalised scores for all 28 brands across all 3 platforms from Run #1, March 2026. Scores are 0–100 normalised within each vertical per platform. Delta = Perplexity minus ChatGPT. A score of 0.0 means the brand was not mentioned in any of the 6 responses in that vertical on that platform.

SEO & Marketing Tools — 12 brands

BrandChatGPTPerplexityGeminiΔ P−GPTArchetype
Semrush90.5100.0100.0+9.5👑 Dominant
Ahrefs100.051.960.7−48.1🧠 AI Memory
Google Search Console61.933.335.7−28.6🧠 AI Memory
Moz47.63.732.1−43.9🩽 Fading
Surfer SEO14.329.614.3+15.3—
Ubersuggest28.611.17.1−17.5—
Screaming Frog23.87.417.9−16.4—
Yoast SEO19.13.73.6−15.3—
Majestic14.30.03.6−14.3—
SE Ranking4.811.17.1+6.3—
Mangools4.811.110.7+6.3—
Rank Math9.53.73.6−5.8—

CRM & Sales — 8 brands

BrandChatGPTPerplexityGeminiΔ P−GPTArchetype
Salesforce100.0100.0100.00.0👑 Dominant
HubSpot74.170.880.0−3.2👑 Dominant
Zoho CRM40.733.336.0−7.4⭐ GEO Outlier
Pipedrive18.533.320.0+14.8—
Freshsales18.58.312.0−10.2—
Insightly14.84.20.0−10.6—
SugarCRM14.80.00.0−14.8—
Keap7.40.00.0−7.4—

AI & LLM Tools — 8 brands

BrandChatGPTPerplexityGeminiΔ P−GPTArchetype
ChatGPT100.0100.0100.00.0👑 Dominant
Copilot43.536.847.8−6.6⭐ GEO Outlier
Jasper39.152.68.7+13.5—
Claude13.063.243.5+50.1🔍 Live Search
Notion AI4.321.10.0+16.7—
Writesonic8.715.88.7+7.1—
Copy.ai13.010.58.7−2.5—
Grammarly13.010.54.3−2.5—

What does the delta mean?

ChatGPT cites brands from its training data — a fixed snapshot of the internet up to its knowledge cutoff, typically one to two years old. Perplexity cites brands from the live web — content indexed and retrievable today. The delta (Perplexity minus ChatGPT) is the gap between those two realities. Negative delta: AI memory is ahead of live web. Positive delta: live web is ahead of training data.

This distinction is the central insight the index was built to surface. A large negative delta means a brand is recommended on historical authority alone. When models next retrain on current web data, that advantage will erode unless the brand is generating current content. A large positive delta means a brand is winning the live web right now — before training data has caught up. That advantage is also temporary; it narrows as models absorb the new content.

Platform parity (delta near zero) is the healthiest long-term position — it means the brand is cited consistently on both historical grounds and current ones. Full delta interpretation guide with threshold ranges →

All 54 queries used in March 2026

The index runs 18 queries per platform — 6 per vertical — sent identically to ChatGPT, Perplexity, and Gemini. The same queries run every month. 6 × 3 verticals × 3 platforms = 54 total query runs per monthly cycle.
SEO & Marketing Tools
  1. What are the best SEO tools in 2026?
  2. Which SEO platform should I use for keyword research?
  3. What is the best tool for backlink analysis?
  4. What are the top tools for technical SEO audits?
  5. Which SEO software is best for content optimisation?
  6. What is the best alternative to Semrush?
CRM & Sales
  1. What is the best CRM for small businesses in 2026?
  2. Which CRM platform should I use for sales teams?
  3. What are the top customer relationship management tools?
  4. What CRM integrates best with email marketing?
  5. Which CRM is best for managing sales pipelines?
  6. What is the best alternative to Salesforce?
AI & LLM Tools
  1. What are the best AI writing tools in 2026?
  2. Which AI assistant is most useful for content creation?
  3. What is the best AI tool for productivity?
  4. Which large language model is best for business use?
  5. What are the top AI tools for marketing teams?
  6. What is the best alternative to ChatGPT?

Queries were selected to reflect genuine evaluation intent — how a user actually compares tools, not brand-specific navigational queries. The panel is fixed month-to-month so scores are directly comparable over time. Any change to the query panel requires a versioned methodology update. Full query selection rationale →

Which SEO tools are fading from AI search in 2026?

The March 2026 data shows Moz (3.7), Yoast SEO (3.7), Screaming Frog (7.4), and Majestic (0.0) all scoring dramatically lower on Perplexity than ChatGPT. These brands are present in AI training data but have weak current web presence. Moz has the largest negative delta in the SEO vertical: 47.6 on ChatGPT, 3.7 on Perplexity, a gap of −43.9.

A brand fading from AI search means two things are happening: the live web is producing less new content that prominently features them, and retrieval-based AI platforms are detecting that absence. Moz was a standard reference in every SEO guide written between 2015 and 2022. ChatGPT learned from that corpus. Perplexity reads the web in 2026. The gap between those two realities is −43.9.

Majestic’s score of 0.0 on Perplexity means it was not mentioned in a single one of the 6 Perplexity SEO responses. It still scores 14.3 on ChatGPT — it exists in training data. But the live web the index monitors in 2026 has moved entirely past it. SugarCRM (0.0 on Perplexity) and Keap (0.0 on Perplexity) show the same pattern in the CRM vertical.

Which brands are dominant across all AI platforms?

In March 2026, ChatGPT, Salesforce, and Semrush were the most consistently cited brands across all three platforms. ChatGPT and Salesforce both scored 100/100/100 — the maximum across all platforms. Semrush scored 90.5/100/100, the only brand where Perplexity and Gemini exceeded ChatGPT. HubSpot scored 74.1/70.8/80.0 — the most consistent second brand in any vertical.

Platform-consistent brands have genuine category ownership: cited because they have earned and maintained live web position over years and continue to generate current content that retrieval platforms find. The distinction between a Dominant Brand and an AI Memory Brand is that Dominant Brands hold their scores across all three platforms. AI Memory Brands have high ChatGPT scores that fall significantly on Perplexity.

Semrush is particularly interesting: it is the only brand in the entire dataset where the live web score (Perplexity: 100.0) exceeds the training data score (ChatGPT: 90.5). This means the web in 2026 has actually warmed to Semrush more than historical training data alone reflected — the mirror image of what Ahrefs is experiencing in the same vertical.

What do the brand archetypes mean?

Archetypes are a diagnostic framework: they describe the pattern of a brand’s scores across the three platforms, not just the absolute level. There are five archetypes in the index. Each one points to a different strategic situation.
👑
Dominant Brand
High, consistent scores across all three platforms.

Cited from both training data and the live web. No AI visibility risk. March 2026 examples: ChatGPT (100/100/100), Salesforce (100/100/100), Semrush (90.5/100/100), HubSpot (74.1/70.8/80.0).

🧠
AI Memory Brand
High ChatGPT score. Significantly lower Perplexity score.

Cited from historical training data. Live web presence is weaker than AI reputation. Requires active GEO investment to defend. March 2026 examples: Ahrefs (delta −48.1), Google Search Console (delta −28.6).

🔍
Live Search Brand
Perplexity score significantly higher than ChatGPT.

Winning on the current web before training data has caught up. Advantage is temporary — narrows as models retrain. March 2026 example: Claude (delta +50.1, ChatGPT: 13.0, Perplexity: 63.2).

🩽
Fading Brand
Low scores across all platforms. Large negative delta.

Losing ground on both training data and live web simultaneously. Requires significant content and citation strategy to recover. March 2026 example: Moz (47.6/3.7/32.1, delta −43.9).

⭐
GEO Outlier
Consistent across platforms, above domain authority expectations.

Scores are balanced and stable — performing above what raw domain authority or market size alone would predict. Suggests effective GEO or content strategy. March 2026 examples: Zoho CRM (40.7/33.3/36.0), Copilot (43.5/36.8/47.8).

Full archetype assignment criteria with thresholds →

How are the scores calculated?

Each brand mention is scored by position: 1st = 5 pts, 2nd = 3 pts, 3rd = 2 pts, 4th+ = 1 pt. A cited URL adds +2. Raw scores are summed across 6 queries per vertical per platform, then normalised to 0–100. The brand with the highest raw total in a vertical on a given platform receives 100. All others scale proportionally.

Scores represent relative citation frequency within this specific query panel — not absolute mention counts or domain authority. A score of 100 means that brand was the most consistently cited in its vertical on that platform. Scores are comparable within a vertical across months, but not across different verticals. Complete methodology with worked examples, normalisation formula, academic references, and known limitations →

View the interactive leaderboard

All 28 brands. Sortable by vertical, platform, delta, and archetype.

Open the index →
Frequently Asked Questions

Questions about the GEO Brand LLM Citation Index

The GEO Brand LLM Citation Index — also called the GEO Brand Citation Index — is a monthly dataset published by Artur Ferreira at The GEO Lab. It measures how frequently 28 brands are cited by ChatGPT, Perplexity, and Gemini when asked 54 standardised evaluation queries across three verticals. It is the first publicly available, regularly updated, methodologically transparent dataset of this kind.
An AI Memory Brand is a brand that scores significantly higher on ChatGPT than on Perplexity. ChatGPT recommends brands based on its training data — a fixed snapshot of the internet from before its knowledge cutoff. Perplexity retrieves the live web in real time. A large negative delta (ChatGPT much higher than Perplexity) means the brand is cited from AI memory, not from current content. This matters because training data eventually updates. When it does, brands with no current web presence will lose their training data advantage too. Full definition →
ChatGPT is a language model trained on a fixed dataset. It cites brands based on how frequently they appeared in that training corpus — which may be one to two years old. Perplexity is a retrieval-augmented generation (RAG) system that searches the live web before generating each response. The gap between their scores tells you whether a brand’s AI visibility is forward-looking (current web presence) or backward-looking (historical training data). Platform definitions →
A score of 0.0 means the brand was not mentioned in any of the 6 Perplexity responses for that vertical’s query set in that monthly run. It is not a rounding error or a data gap — the brand genuinely did not surface. In March 2026, SugarCRM, Majestic, and Keap all scored 0.0 on Perplexity. SugarCRM still scored 14.8 on ChatGPT, which means it exists in training data — but the live web the index monitors has moved past it entirely.
Monthly. The same fixed panel of 54 queries runs every month. Results are normalised within each vertical and published alongside the monthly report. Scores from different months are directly comparable because the query panel does not change. Any change to the methodology requires a versioned update in the methodology changelog.
No. Scores are normalised within each vertical separately. A score of 50 in SEO & Marketing and a score of 50 in CRM & Sales are not equivalent — each is relative to the top brand in that vertical on that platform. Cross-vertical comparisons should use archetypes and delta patterns, not raw scores. Normalisation methodology →
A large negative delta is a live web presence problem, not an AI problem. The fix is content that gets indexed, cited, and linked today: original data, placements in comparison roundups, third-party mentions in category articles. The full strategy is covered in What is GEO? and the GEO Lab experiment series starting at Experiment 001.
The index tracks a curated panel selected for category relevance. Inclusion is based on category fit, not payment. Submit your brand for consideration at thegeolab.net/geo-brand-citation-index/. The panel expands monthly as new verticals and brands are added.
The GEO Brand Citation Index was built by Artur Ferreira, founder of The GEO Lab. It was built because brands had no transparent, repeatable way to measure their AI citation visibility — only anecdote and guesswork. The index has run monthly since March 2026. The GEO Lab publishes research, experiments, and data on Generative Engine Optimisation: the practice of optimising content and brand presence for AI-driven discovery.