--- title: "Across 88 Published AI-Citation Leaderboards, the Analyst Report Never Finishes First" description: "Today's Machine Relations Index classifies citations by source type across 88 published category-and-question segments. Gartner, Forrester, McKinsey and every other analyst or consulting domain combined lead zero of them. Their best showing is third place, in one segment." canonical: https://authoritytech.io/curated/analyst-relations-never-wins-ai-citation-2026 last-updated: 2026-09-25 --- # Across 88 Published AI-Citation Leaderboards, the Analyst Report Never Finishes First Today's Machine Relations Index classifies citations by source type across 88 published category-and-question segments. Gartner, Forrester, McKinsey and every other analyst or consulting domain combined lead zero of them. Their best showing is third place, in one segment. Canonical URL: https://authoritytech.io/curated/analyst-relations-never-wins-ai-citation-2026 Published: 2026-09-25 Author: Jaxon Parrott Tags: Afternoon Brief, AI Search & Discovery, Measurement A marketing leader weighing a Gartner or Forrester analyst-relations program against a content or PR budget line is usually asking the wrong first question. The pitch for the analyst program is enterprise credibility with human buyers and analysts, and that pitch stands on its own. The question this brief answers is narrower: does the same spend also buy citations when a buyer asks an AI answer engine the question instead of a human analyst? Today's [Machine Relations Index](https://machinerelations.ai/index) release gives a direct, checkable answer, and it is no. ## What was measured Release `mri_score_v2.0+2026-09-25+04fcb7fb8f29`, generated 2026-09-25, observation window 2026-05-10 through 2026-09-25, 132 observed days, read whole from [machinerelations.ai/data/machine-relations-index.json](https://machinerelations.ai/data/machine-relations-index.json) in an isolated compute environment. The Index scores 23,872 cited domains across six engines — ChatGPT, Claude, Gemini, Google AI Mode, Google AI Overviews and Perplexity — against 25 subject categories, each paired with one of six buyer-question shapes: best tools, how buyers choose, is it worth it, problem-first research, top lists, and head-to-head comparisons. Every cited domain also carries a Source Role label from the Index's own declared taxonomy: editorial publication, vendor-owned, market and company database, academic and government, community and social, analyst and consulting research, search or media platform, wire and press-release distribution, or an unclassified "other observed source" layer that this brief excludes throughout, the same way [today's category-capture study](https://machinerelations.ai/research/source-class-capture-by-category-2026) does — it is the majority of total citation volume and is not a finding about any named class. A category-and-shape pair enters this brief only once it clears the release's evidence floor of at least 10 observed answer runs across at least 7 distinct run dates and carries at least one classified-source citation. 88 segments across the six buyer shapes clear that bar today. ## The result Group each segment's classified citations by Source Role and find the leader. Across the 88 segments, the leader is an editorial publication in 38, a vendor-owned domain in 29, a community or social platform in 10, an academic or government source in 9, and a market or company database in 2. Analyst and consulting research — the class that holds [gartner.com](https://www.gartner.com), [mckinsey.com](https://www.mckinsey.com), [deloitte.com](https://www.deloitte.com), [pwc.com](https://www.pwc.com) and [ey.com](https://www.ey.com) — leads zero. It is not close in the segments where it appears at all, and it is absent from most of them. Analyst and consulting sources carry any classified citations in 45 of the 88 segments; in the other 43, the class contributes nothing. Where it does appear, its best result across the entire release is third place, in Enterprise Software's top-lists segment, where its citations are a fifth of that segment's classified total — behind market-and-company-database sources and vendor-owned sources, both ahead of it. Its next-best result, in Fintech's top-lists segment, is a rank of second, at roughly a sixth of the classified total, still behind the segment's leader. Everywhere else it measures, it ranks fourth or lower, and in 15 of the 45 segments where it shows up at all, it is dead last among the classified roles present. This holds even though the class's flagship domain is not thin evidence. [gartner.com](https://www.gartner.com) is the single most-cited analyst-and-consulting domain in the release, graded confidence A — the release's highest tier, meaning it clears the deepest evidence bar the Index sets. One well-evidenced domain carrying most of a class's citation volume, and the class it anchors still does not win a single published buyer-question leaderboard. ## Why this is not the same finding as "no source wins every shape" [A separate brief published on this release](https://authoritytech.io/curated/question-shape-decides-ai-citation-winner-2026) established that no single domain or source type sweeps every buyer question in a category — the winner changes with the shape of the question. That finding is about which class wins where. This one is about a class that does not win anywhere, in any category, on any shape, in the current release. Those are different claims. A brand could read the first finding and conclude every channel has its shape to win; this finding says one channel measured here does not currently have one, at least not one large enough to clear a published leaderboard's top spot. ## What a comparable dollar buys elsewhere This brief does not put a price on an analyst-relations program — vendor spend varies too widely by firm tier and company size to state one number responsibly, and Gartner's own [Office of the Ombuds](https://www.gartner.com/en/about/ombuds/news/cost-included-gartner-research-answer-simple-nothing) is explicit that inclusion in its research itself is not for sale at any price: "What does it cost to be included in Gartner research? The answer is simple: nothing." The [discipline of analyst relations](https://en.wikipedia.org/wiki/Analyst_relations), as [the industry's own trade body defines it](https://analystrelations.org/the-ar-job-description/), is built around briefings, inquiries and research-participation time, not a line item you can look up. What is comparable is the source role that a program like that feeds: analyst and consulting research, measured here on its own terms, at the same evidence bar as [g2.com](https://www.g2.com) and [crunchbase.com](https://www.crunchbase.com) (market and company database, 2 leaderboard wins of 88), [reddit.com](https://www.reddit.com) and [linkedin.com](https://www.linkedin.com) (community and social, 10 wins), and [forbes.com](https://www.forbes.com) and [medium.com](https://medium.com) (editorial publication, 38 wins). ## What this does not say It does not say analyst relations has no value — briefing analysts, shaping inclusion criteria, and being named in a Magic Quadrant or Wave reach human buyers and procurement committees this Index does not measure at all, and that reach was never this Index's subject. It does not say a specific analyst report can never be cited; [nih.gov](https://www.nih.gov) and [arxiv.org](https://arxiv.org) show that a specific, well-evidenced research domain can lead its own class elsewhere in the release. It says that as a class, at the current release, across every category and buyer-question shape with enough evidence to publish, analyst and consulting research has not produced a single #1 finish, and it is worth knowing that before a marketing budget assumes an analyst program does double duty as an AI-citation strategy. It also does not extend past this release or this taxonomy. A future release, a different classification of a borderline domain, or a segment that has not yet cleared the evidence floor could change the picture, and the next release is the way to check whether it has. ## Sources and method [Machine Relations Index](https://machinerelations.ai/index) public view, contract `machine_relations_index_public_view_v2.0`, methodology version `mri_score_v2.0`, release `mri_score_v2.0+2026-09-25+04fcb7fb8f29`, generated 2026-09-25, read whole from [machinerelations.ai/data/machine-relations-index.json](https://machinerelations.ai/data/machine-relations-index.json) and cross-checked against the [release manifest](https://machinerelations.ai/data/mri-release-manifest.json) in an isolated compute environment on 2026-09-25. Window 2026-05-10 to 2026-09-25, 132 observed days. Evidence floor for a published segment: at least 10 observed answer runs across at least 7 distinct run dates. Source Role definitions and taxonomy are the release's own, explained at [the Source Role glossary entry](https://machinerelations.ai/glossary/source-role). Per-domain source-role classification and per-stratum citation counts are read directly from the release's own fields, not re-derived. "Other observed source," the unclassified majority of total citation volume, is excluded from every denominator in this brief, as it is in the release's own category-capture study cited above. The observation window includes a period, 2026-08-15 through 2026-09-24, when two of the six engines — Google AI Mode and Google AI Overviews — had a collector-side collection gap on parts of the window; [Perplexity](https://docs.perplexity.ai), ChatGPT, Claude and Gemini were collected continuously throughout. That gap can only ever reduce the total run volume a segment accumulates during the window; it gives no engine or source class a directional advantage in which domain within a segment gets cited most, so it does not change which class leads or fails to lead a segment. This brief makes no claim about any single engine's coverage of any source class, and no comparison across engines; it groups citations by source role across all six engines combined. Analyst-relations program mechanics referenced above: [Wikipedia's Analyst Relations entry](https://en.wikipedia.org/wiki/Analyst_relations); the [International Institute of Analyst Relations' AR job description](https://analystrelations.org/the-ar-job-description/); Gartner's own [Office of the Ombuds statement on research cost](https://www.gartner.com/en/about/ombuds/news/cost-included-gartner-research-answer-simple-nothing). ## Links - [Curated Index](https://authoritytech.io/curated.md) - [Home](https://authoritytech.io/index.md)