Gane Index · November 2026 Edition

What AI models recommend for business software

Every month, fourteen AI models from twelve labs are asked the same buying questions, six ways, for small business, mid-market and enterprise buyers. Each category shows the products they choose most consistently, with their reasons and reservations in their own words.

This edition

  • 363categories
  • 23,710products
  • 9,130vendors
  • 14models

Across every edition

  • 249,000AI answers read and judged
  • 125Mwords in those answers
  • 1.38Mjudge labels on those answers
  • 2.78Msource links traced, across 75,000 sites
How the Index works
  • Six framings. Each buying question is asked six ways: direct, paraphrase, comparative, budget-constrained, scale-constrained and negative. Share counts the four that ask for a pick (direct, paraphrase, budget and scale); the comparative and negative answers are read for comparisons and warnings.
  • Three buyer sizes. Every question is asked for a small business, a mid-market and an enterprise buyer, and each size is its own standing, never added together.
  • A disclosed judge. Every answer is read by a judge model, named on each method page, which labels each product named and quotes the sentence the label rests on.
  • The noise floor. The same questions are asked twice with nothing changed. How far a leader's share moves then is the floor, and a smaller change between editions is reported as within noise.
  • Monthly. A new edition every month, rescored in full. The next edition runs December 1, 2026.

The method for each index: GTM, IT, HR, Finance.

Browse GTM categories

Browse IT categories

Browse HR categories

Browse Finance categories