How to Build AI Citations — Whiteboard Friday

Step 2: Analyze the citations for each prompt

Once you’ve got a short list of prompts, let’s say you’ve got up to 20 prompts that you’re working with to begin with, you can then input those manually into ChatGPT or the AI platforms that you want to show up for and start analyzing the citations that get given in those responses.

Most of the time, there’s going to be quite a few citations. So I would recommend creating a full list of citations in a spreadsheet, with the URLs marked. And then you can start to see the frequency of certain topics and of certain publications.

Step 3: Perform outreach for inclusion in those cited publications

Then, once you’ve got those, you can outreach to the journalists, to the editors who wrote those articles, to try and get your business mentioned within the article as well.

And what’s most important here is not getting a link in the article, the way that we think of link building for traditional Google SEO. It’s actually just having your brand mentioned in the context that you care about, that feels like the way you want your business or your client’s business to be understood by AI.

So don’t worry too much about the link. Think much more about the mention and the context.

But this is kind of a reverse engineering strategy at the moment. We’re actually seeing what’s already been cited and then trying to gain mentions in those citations. We can do this the other way around. 

BONUS: Do proactive outreach to gain additional mentions

So the bonus example is thinking about proactive outreach. Once you’ve got your whole list of topics and citations, you can get an idea of the websites that get cited really frequently. You might start to see the same domains coming up and the topics that come up frequently as well.

For example, if you’re a business that actually sells CRMs to other small businesses, you might be searching for topics like the best CRMs for small businesses or the best CRMs for small sales teams.

In those types of citations, you might see a lot of “best of ” articles. That gives you a way to proactively decide what you actually want to outreach for, and that might be trying to get links in those articles in the short term. But next year, you might be thinking about publishing 2026 articles or 2027 articles on similar “best of” topics, knowing they’re quite likely to be referenced by AI.

One of the brilliant things about the proactive outreach is that AI loves to cite fresh content. So anything that gets published more recently has more potential to be cited, especially if you know it’s the kind of website that AI often cites from your original reverse engineering strategy.

If you have any more questions about AI citation building and how it works, then I’m Charlie Marchant from Exposure Ninja. You can find me on LinkedIn anytime, and I’m more than happy to have a chat. Have a fantastic rest of your Friday.