# Machine Relations Index (MRI)

The Machine Relations Index (MRI) is a public source-behavior dataset that tracks which root domains answer engines — ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode — cite when responding to B2B buyer-intent questions. It classifies every observed source by deterministic source-role rules, measures engine coverage and vertical spread, and publishes the full cited-domain set with query evidence.

Canonical URL: https://machinerelations.ai/glossary/machine-relations-index
Category: metrics
Attribution: Coined by Jaxon Parrott.

## What the MRI measures

The Machine Relations Index monitors how answer engines select sources when they respond to commercial research questions. It captures every root domain cited across a monitored B2B buyer-query set, classifies each domain by its source function, and tracks citation behavior across five answer engines over a rolling observation window.

The MRI is not a ranking of brands or a quality score. It is a behavioral map of the source layer that machines use when they construct answers to buyer-intent queries. A domain appears in the MRI because at least one engine cited it in at least one observed query — nothing more.

## Source-role taxonomy

Every domain in the MRI is classified by deterministic rules into one of nine source roles based on its function in the citation ecosystem:

- **Editorial publication** — news and trade media (e.g. TechCrunch, Fierce Healthcare)
- **Analyst and consulting research** — advisory firms (e.g. Gartner, Forrester, Deloitte)
- **Market and company database** — data platforms (e.g. Crunchbase, G2, PitchBook)
- **Academic and government source** — institutional knowledge (e.g. NIH, Stanford, Wikipedia)
- **Community and social platform** — user-generated surfaces (e.g. Reddit, LinkedIn, GitHub)
- **Wire and press-release distribution** — syndication networks (e.g. PR Newswire, Business Wire)
- **Search or media platform** — discovery infrastructure (e.g. YouTube, Google)
- **Vendor-owned source** — company domains explicitly identified (e.g. Salesforce, Stripe)
- **Other observed source** — long-tail domains not yet classified by deterministic rules

The "Other observed source" bucket is deliberately large. The MRI does not overclaim classification: domains remain uncategorized until deterministic evidence supports a role assignment. This preserves research integrity over cosmetic completeness.

## Dimensions of the index

The MRI tracks four dimensions for every cited domain:

1. **Citation volume** — total citation events in the observation window, used as the primary ranking metric
2. **Engine coverage** — which of the five monitored engines cite this source, revealing cross-engine trust patterns
3. **Vertical spread** — how many B2B industry verticals the source appears in, indicating breadth of machine trust
4. **Query evidence** — sample prompts that triggered citations, making the relationship between query intent and source selection legible

## Why it matters for Machine Relations

Machine Relations is the discipline of becoming legible, credible, and citable to machine-mediated discovery systems. The MRI is the empirical foundation: it shows what the source layer actually looks like, which sources engines trust across contexts, and where citation authority concentrates or disperses.

For practitioners, the MRI answers a direct question: when a buyer asks an AI engine about your category, which sources does the engine reach for — and is yours among them?

The MRI was coined by Jaxon Parrott and is maintained as a public research artifact at [machinerelations.ai/index](https://machinerelations.ai/index). Machine-readable versions are available as [JSON](https://machinerelations.ai/data/machine-relations-index.json) and [Markdown](https://machinerelations.ai/machine-relations-index.md).

## Sources

- https://machinerelations.ai/index
- https://machinerelations.ai/data/machine-relations-index.json
- https://machinerelations.ai/research/machine-relations-stack-five-layers
- https://medium.com/authoritytech/machine-relations-explained-76e9f174377c
- https://machinerelations.ai/glossary/machine-relations