Get onto the pages AI already cites.
When an LLM answers a question in your category, it builds the answer from a handful of pages. If you're not on those pages, you're not in the answer. Citation outreach finds them, prompt by prompt, from the answers we already collect for you, and turns them into outreach that gets you added.
Find the gapsPitch the sourcesGet cited
Illustration · your targets come from your own tracked prompts
How it works. Monitoring you already have, pointed at the pages that decide the answer.
01
Your prompts get asked
The questions you track are put to ChatGPT, Claude, Gemini and Perplexity on a schedule. Every answer is kept in full, with the pages it cited.
02
The gaps become targets
We look for answers that cite somebody else and not you. The cited page is the target, and it's grouped by the prompt that surfaced it — so a target always arrives with the question it came from and the LLMs that gave it.
03
The list gets worked
Targets go to our outreach partner, who contacts the publisher to get your brand added to the page. The pitch is written against the prompt, not sent as a template.
A target is a page and the question that surfaced it. Never just a domain on a list.
Every row says which question it came from, which LLMs answered it that way, and who is on the page instead of you, which is what makes a pitch writable.
| Prompt | Page cited | Named there instead of you | Cited by |
|---|---|---|---|
| best budgeting app | techradar.com/best-budgeting-apps | Four competitors | ChatGPT, Perplexity |
| budgeting app for couples | nerdwallet.com/article/finance/best-budget-apps | Two competitors | Claude, Gemini, Perplexity |
| ynab alternative | forbes.com/advisor/banking/best-budgeting-apps | One competitor | ChatGPT |
Illustration · sample rows
What it doesn't do. Worth saying outright, because this sits next to a category that promises all of it.
No guaranteed placement
The publisher decides what goes on their page. Outreach gets you asked for, properly and in context; it can't promise a yes, and anyone promising one is selling something else.
No bought links
This is editorial outreach — a pitch to be included on a page AI already trusts. Nothing here buys a link or places paid content.
Not a content tool
It doesn't write your pages. The targets are other people's pages, which is the whole point: the LLM is citing them instead of you.
Not a link-authority play
Targets aren't chosen by domain score. They're chosen because an LLM cited that exact page to answer a question your buyers ask.
Frequently asked questions
Your own tracked prompts. We ask them across the LLMs on a schedule and keep every answer with its citations, so a target is a page an LLM really did cite when it answered a question you chose to track. Nothing is scraped from someone else's conversations, and nothing is modelled.
A page cited in an answer where your brand isn't cited. Your own pages are excluded, and competitor-owned pages are separated out — you're not going to get added to a rival's site. What's left is third-party pages: the roundups, comparisons and buyer's guides on blogs and publications.
It scales with the prompts you track and how often you're missing from the answers. A few hundred prompts typically surfaces thousands of prompt-and-page pairs, which is why they arrive as a worked list rather than a report to read.
They're identified separately, because a community thread is a different kind of approach from a publisher's roundup, and the same outreach doesn't suit both. You choose whether they're in scope.
An outreach partner we hand the list to, working from the prompt each target came from. You see the same targets we send.
It's set up with your account team rather than switched on in settings, because the outreach side involves people, not just a connection. Email us and we'll walk you through what your prompts would produce.