If you ask ChatGPT "what's the best link shortener for marketing teams?" right now, it will recommend a handful of tools by name. Some of those tools are your direct competitors. Yours may or may not be on the list.
That recommendation happens not because those companies paid for it or submitted their names somewhere. It happens because ChatGPT has seen enough credible, consistent content about those tools that it trusts them as legitimate answers to that question.
This is the new category of SEO that nobody has a perfect playbook for yet — and that's exactly why early movers have an advantage.
How ChatGPT decides what to recommend
ChatGPT doesn't rank things the same way Google does. There's no PageRank equivalent, no domain authority score it checks. The process is closer to what a very well-read expert would do if you asked them for a recommendation after they'd read millions of articles, reviews, comparisons, and case studies on the topic.
The brands that show up consistently are the ones that:
- Have been written about frequently and positively across multiple independent sources
- Have clear, well-structured content on their own site that directly answers the questions users ask
- Are associated with specific, well-understood use cases rather than vague "all-in-one" positioning
- Appear in comparison articles, "alternatives to" lists, and category roundups across the web
Notice that none of those are things you can fake or shortcut. They're the byproduct of actually building a credible product and then making sure the public record reflects that credibility.
The two ChatGPT contexts to optimise for separately
Context 1: Base model recommendations (no browsing)
When a user asks ChatGPT a question without enabling browsing, the response is drawn entirely from training data — content the model was trained on before its cutoff date. To appear here, you need to have generated enough web presence before that cutoff: reviews, comparisons, forum discussions, case studies, and mentions in reputable publications.
This is a long game. It's also why established brands show up here more reliably than newer ones.
Context 2: ChatGPT Browse and Search (live web)
When ChatGPT's browsing mode is active, it fetches real-time content from the web and cites sources. This is the context where SEO work you do today can affect what ChatGPT recommends this month.
Browsing mode essentially makes ChatGPT behave like a very smart research assistant running a Google search on your behalf. The pages it tends to cite from are:
- Pages with clear, direct answers near the top
- Pages from sites with consistent publishing on the topic
- Pages that have been cited or linked to by other credible sources
This is where your content library matters enormously. If you have 30 guides on marketing attribution and link tracking, ChatGPT is far more likely to cite your UTM guide than a single standalone page from a site with no topical depth. Topical authority is as important here as it is in traditional SEO.
Practical steps to increase your ChatGPT citation rate
Step 1: Audit what ChatGPT currently says about your category
Open ChatGPT Browse and ask the questions your customers actually ask:
- "What's the best link tracking tool for small businesses?"
- "How do I track UTM parameters without Google Analytics?"
- "Which URL shortener is best for branded links?"
Document which brands appear in the answers and which sources get cited. The cited pages are your benchmarks. They're doing something well that your pages aren't.
Step 2: Structure your key pages around direct answers
ChatGPT pulls the most citable content from pages where the answer to a question appears in the first 1-2 paragraphs under a clear heading. Not buried in the middle. Not at the end.
Rewrite your most important landing pages so the first thing under each section heading is a direct, factual statement. For a tool page: "The [Tool Name] is a [what it does] that [key benefit] — used by [audience] to [outcome]." That format is much easier for an AI system to extract and use in a synthesised answer than a paragraph that starts with "In today's digital landscape..."
Step 3: Publish content that fills citation gaps
Whatever questions your category is associated with — if you don't have a dedicated page that directly answers it, you will not get cited for it. Identify the 10 most common questions your customers ask, and make sure each one has a page on your site that answers it better than any competitor.
This is exactly what the SEO calendar approach is built around — not just ranking for keywords, but owning the questions that define your category.
Step 4: Get cited externally
The pages ChatGPT cites most reliably are pages that other sites link to. Traditional link building is not dead — it's now doubly important because backlinks are a proxy for the external credibility signals that both Google and AI systems use to decide what to trust.
Guest posts, partnerships, PR mentions, and appearing in comparison articles ("X vs Y" or "best tools for Z") all contribute to the pool of content where your brand is mentioned and linked to. Each of those mentions increases the probability that an AI system, reading that content, will consider your brand a legitimate answer to a category question.
Step 5: Track whether it's working
Measure ChatGPT-referred traffic in your analytics (the ChatGPT referral traffic guide explains how to do this), and manually check your target queries in ChatGPT Browse every few weeks. The goal is to see your pages appear in citations for the specific questions that drive purchase intent in your category.
When they do start appearing, make sure those cited pages have visible CTAs that convert the visit into something measurable — whether that's a free tool, a guide download, or a signup. A user who clicked a ChatGPT citation is already half-convinced. Don't make them dig for the next step.
Your Link Trust Predictor and UTM Builder are exactly the type of no-friction, instant-utility tools that AI systems confidently recommend when users ask practical questions. Make sure the copy on those pages uses the exact language users search with — not the marketing language your team prefers internally.