> ## Content Index
> Fetch the complete content index at: https://ai-visibility-engineering.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# Getting Cited by AI: An On-Site and Off-Site Checklist
- URL: https://ai-visibility-engineering.ghost.io/getting-cited-by-ai-an-on-site-and-off-site-checklist/
- Published: 2026-07-27T16:38:45.000Z
- Updated: 2026-07-27T16:47:23.000Z
- Author: Derek Iwasiuk
- Tags: Playbooks

Getting cited by an AI model isn't one task, it's two. There's the on-site work: making a brand legible to a model in the first place. And there's the off-site work: giving the model outside reasons to trust and repeat what it found. Most teams only do one. Both are required.

## On-site: make yourself easy to extract

This is about removing every reason a model has to guess.

- Schema markup on the pages that matter: Organization, Product, FAQ, and Article schema so structured facts are machine-readable, not just visible.
- One clear, current statement of what the business does and who it's for, without relying on tone or branding to imply it.
- An FAQ section that states the questions people actually ask a model, in plain language, with direct answers.
- Consistent facts across every page. Pricing, positioning, and claims that contradict each other across the site erode a model's confidence in all of it.
- A crawlable, fast, text-first version of key pages. Content buried behind heavy JavaScript is content a model may never see.

## Off-site: give the model outside reasons to trust it

A model rarely relies on a single source. It looks for agreement across independent ones.

- Earned citations from publications and sites the model already treats as authoritative in the category.
- Placement in the comparison posts and listicles models pull from directly, the “best X for Y” roundups.
- A complete, accurate presence on the review and data platforms relevant to the industry, G2, Capterra, Crunchbase, Wikipedia, and similar.
- Expert quotes and guest contributions that put the brand's name next to a specific claim or specific expertise.
- A presence in the forums and Q&A communities models increasingly draw from, where real users already discuss the category.

## How to tell if it's working

Ask the models directly. Prompt ChatGPT, Gemini, Copilot, and Perplexity with the questions a real prospect would ask in the category, and check whether the brand shows up, whether the facts are right, and which sources get credited when it does. That answer is the actual scoreboard. Rankings and traffic are a proxy for it, not a replacement.

This is the kind of practical, tactical ground this newsletter will keep covering, in partnership with SearchTides.