Use transparent AI-search evidence other teams can inspect and cite.

Original research earns attention when the question, sample, method, exclusions and limits are visible. These resources separate observed provider answers from vendor claims and planning assumptions so a reader can reuse the finding without overstating what the dataset proves.

How to use this collection

Decision principle

Publish the denominator

Every finding needs its questions, providers, successful observations, time window and exclusions. A percentage without scope is not reusable evidence.

Decision principle

Distinguish source types

Official vendor material can establish a product claim. It cannot independently prove that the product caused a ranking or customer outcome.

Decision principle

Make the artifact reusable

Clear tables, methods and downloadable working material help analysts, journalists and buyers apply the research to their own decision.

Start with these guides

Buyer guide

Every guide in this collection

Buyer guide

See what Eli would fix on your website first

Free account. No card.

Try Eli for free

Use transparent AI-search evidence other teams can inspect and cite.

How to use this collection

Publish the denominator

Distinguish source types

Make the artifact reusable

Start with these guides

What 31 premium SEO agencies agree on about AI search.

What 20 AI-search monitoring and content platforms actually own.

What a 2,000-question B2B AI-search market map is designed to reveal.

AI-search monitoring versus execution: a verification matrix.

Every guide in this collection

How to study citation overlap across ChatGPT, Gemini, Claude and Perplexity.

What 193 AI-cited Reddit threads can teach an AEO team.

Methodology for a B2B AI-crawler access benchmark.

How to analyze the source types cited in B2B AI-search answers.

See what Eli would fix on your website first