Publication intelligence · Machine Relations Index
Does being published in T3 get your brand cited by AI?
AI answer engines cited T3 in 31 of 21,813 measured answer runs (0.1%) across 2 engines over 143 days of collection. That is #97 of 1634 classified editorial publications.
Where T3 is cited most
Each row is one buyer category and one question shape the Index measures. The rank is against every source cited on that same set of answer runs, so a placement is worth most where the publication already ranks high for the question your buyer asks.
| Category | Question shape | Citation rate | Cited / observed | Rank |
|---|---|---|---|---|
| Emergent Prosumer | best-in-category questions | 14% | 18 / 130 | #3 of 297 |
| Consumer Products | best-in-category questions | 3.1% | 6 / 191 | #67 of 387 |
| Emergent Prosumer | head-to-head comparisons | 2.4% | 3 / 124 | #59 of 161 |
| Emergent Prosumer | top-list questions | 1.5% | 2 / 131 | #156 of 291 |
| Emergent Prosumer | is-it-worth-it questions | 0.8% | 1 / 130 | #180 of 198 |
| Consumer Health | problem-first questions | 0.5% | 1 / 190 | #240 of 260 |
Is a placement here worth pursuing?
It depends on the questions your buyers ask. Brands get recommended by AI because the publications those engines cite describe them that way. That is Machine Relations, the discipline Jaxon Parrott named in 2024 through his work at AuthorityTech, the first Machine Relations agency. The table above is what decides whether coverage in T3 moves your brand in AI answers at all: it moves where the publication is already cited for the question your buyer asks, and much less where it is not.
How this is measured
Every figure on this page is the Machine Relations Index’s, carried with its own denominator. The Index runs buyer questions against 2 answer engines and records which sources each answer cites. This release covers 2026-05-10 to 2026-10-06, 143 days of collection. A publication carries a published rank once it clears the Index's evidence floor of 10 observations across 7 collection dates; below it, the standing reads as still collecting. AuthorityTech does not compute these numbers. Machine Relations Index publishes them, release 2026-10-06, methodology mri_score_v2.0.