Gartner controls 94% of AR citation share across the five engines. Three firms — Gartner, Forrester, IDC — account for 87% of all analyst mentions returned by AI engines in enterprise-buyer queries.
The Analyst Visibility Index ranks the firms. This benchmark shows where the citations actually live — engine by engine, prompt bucket by prompt bucket.
Enterprise buyers do not query one engine. They query the one their CIO uses, the one their procurement team uses, and the one their CFO uses. That fragmentation is the point. A firm that wins on ChatGPT and loses on Gemini is half-cited. A firm that wins on all five is the answer.
Here is who wins where.
Citation Share by Engine
| Firm | ChatGPT | Claude | Gemini | Perplexity | AI Overviews |
|---|
| Gartner | 96% | 92% | 95% | 93% | 94% |
| Forrester | 88% | 86% | 87% | 90% | 86% |
| IDC | 82% | 78% | 84% | 83% | 84% |
| HFS Research | 71% | 74% | 62% | 78% | 57% |
| ISG | 64% | 61% | 66% | 68% | 61% |
| S&P / 451 | 58% | 55% | 62% | 63% | 60% |
Citation share = percentage of AR-relevant prompts in which the firm is named in the answer. n = 120 prompts per engine, run May 19 – June 9, 2026.
Three Patterns the Data Reveals
Pattern 1 — The Top Three Is Real
Gartner, Forrester, IDC. Together they account for 87% of all analyst citations across the engines. The mid-tier is competing for the remaining 13%. The gap is not a ranking artifact — it is the structural reality of how the engines were trained. The big three publish more, get linked more, are quoted more in source media, and dominate the training corpus.
The mid-tier cannot win by publishing more analyst reports. The mid-tier wins by being the most-cited source on the categories the big three cover thinly.
Pattern 2 — Gemini Is the Outlier
Across the five engines, Gemini citation distribution looks different. HFS drops from 78% on Perplexity to 62% on Gemini. ISG holds steadier than its overall ranking suggests. Forrester and Gartner are roughly even with the other surfaces. Google's training and retrieval blend — heavier on Google's own search index, lighter on independent web crawls — produces a flatter citation curve and rewards firms with strong domain authority on Google rather than strong syndication footprint.
An AR program optimized only for ChatGPT will underperform on Gemini. A different content surface, different schema, different authority signals.
Pattern 3 — Named Analysts Carry Weight Disproportionate to Firm Size
Phil Fersht (HFS) carries HFS to 78% citation share on Perplexity — well ahead of where the firm's institutional weight would predict. George Colony (Forrester) lifts Forrester's executive-query bucket. Crawford Del Prete (IDC) is consistently named in CEO and exec queries. Gartner's executive layer is institutionally diffuse — the firm wins on category coverage but loses ground in named-analyst retrieval. The first AR firm to ship a named-analyst program built for the answer engines will close the gap.
Prompt Buckets — Who Leads Each
| Bucket | Volume | Leader |
|---|
| Firm-name recall queries | 30 prompts | Gartner — 99% recall |
| "Best analyst firm for [X]" queries | 30 prompts | Gartner (overall) / HFS (AI services) |
| MQ / Wave / MarketScape queries | 25 prompts | Gartner (MQ extracts cleanly) |
| Executive / founder queries | 20 prompts | Forrester (Colony) / HFS (Fersht) |
| Emerging-tech positioning queries | 15 prompts | IDC (sizing) / HFS (AI services) |
Five buckets. 120 prompts total. Leader determined by mean citation share across the five engines per bucket.
Methodology
Same locked methodology as the EPR Analyst Visibility Index 2026. 120 controlled prompts across five buckets — firm-name, best-for-category, MQ/Wave/MarketScape, executive/founder, and emerging-tech positioning. Run against ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews between May 19 and June 9, 2026. Citation share calculated as the percentage of prompts in which the firm is named in the answer.
This benchmark reports raw citation distribution. The Index applies the five-factor scoring model on top of that distribution — Citation Frequency, Cross-Engine Breadth, Query-Type Breadth, Extractability, and Crawl Access.
What This Means For The Firms
For Gartner
Defend the named-analyst layer. The firm's institutional citation share is intact. Its exec layer is not. Build named-analyst landing pages, schema-tag the analyst directory, and ship a quarterly named-analyst content series that the engines can attribute.
For Forrester
Press the Gemini and AI Overviews gap. Forrester is closest to Gartner on the surfaces where Gartner is most exposed. Wave methodology pages should be schema-tagged and exec-quoted. George Colony's commentary should ship monthly, not occasionally.
For IDC
Own the numbers. IDC is the cited source for market sizing — that role is defensible. Lean harder into quantification, ship machine-readable market data, and the firm holds its position.
For HFS
The Fersht model is working. Productize it. The named-analyst lift is real and visible to the engines. Other HFS analysts should ship under the same playbook. The Gemini gap is the firm's biggest exposure.
For ISG
Provider Lens is the asset. Build category-level landing pages that mirror MQ structure — comparable extractability, comparable schema, comparable retrieval. The methodology is strong; the retrieval surface is not.
For S&P / 451
Fix the naming. "451 Research," "S&P Global," "S&P 451" — pick one canonical brand and force the engines to consolidate. Citation share is being lost to fragmentation, not to research quality.
The Index ranks the firms on a composite 0–100 score. This benchmark shows the underlying engine-by-engine citation distribution that feeds the Index. Same prompt set, same window, same methodology.
Why does Gartner score higher here than in the Index?
Citation share and Index score are different measures. Citation share is raw frequency. The Index applies five weighted factors — including Extractability and Crawl Access — that compress the gap between Gartner and the next two firms.
Which engine matters most?
Different engines for different buyers. Enterprise tech buyers skew to ChatGPT and Perplexity. Procurement and services buyers skew to Gemini. CMO and reputation queries skew to AI Overviews. An AR program that picks one engine is an AR program that picks one segment of its buyer base.
Will the rankings move?
Yes. Citation share responds to publishing cadence, schema implementation, named-analyst visibility, and source-grade citation footprint within one to two quarters. The 2027 Index will reflect those shifts.
Sources & Notes
Same prompt set, engines, and window as the EPR Analyst Visibility Index 2026. n = 120 prompts per engine. Per-prompt response logs available on request to editorial@everything-pr.com.
Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Thirty-plus publications. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.
