And no, this isn't just an SEO rebrand.

One of the best questions I've ever been asked in my career is: who is this press release trying to reach — customers, investors, or employees?

A fellow PR person who was working with the partner of one of my clients on a launch asked it in our first meeting, and I've stolen it so many times since. It's a reminder that press releases — or really any public communication in general — have one of three audiences. Sometimes, yes, it's a combination of two. Rarely is it ever three.

But now, I'm adding a fourth audience. Computers. And by computers, of course, I mean AI and LLMs.

So, here's another question: what do the computers think about your company? Or even, do the computers know about your company at all?

It's a question I never thought I'd ask, but here we are. And it's changing the way I think about the first question. The answer is no longer just customers, investors, or employees. It's also the systems that are quietly forming opinions about your brand on their behalf.

Unlike customers who don't read quarterly reports or employees who skip the weekly customer newsletters, LLMs don't miss things. They see it all, and use it to influence people asking questions in places you'll never see.

Why this matters (and why now)

AI search is still a small slice of total search volume — maybe 1% today. I know what you're thinking: 1% doesn't justify a strategy shift.

But here's a better way to think about it. We are essentially in 1999 Google, right before incumbents got toppled and the people who figured out SEO early built lasting advantages. The companies that start optimizing for AI search now are setting up to own that real estate for years. The ones who wait until it's "significant enough to matter" are going to spend a lot of money catching up.

How AI finds your company

In order to understand what kind of content drives AI visibility, you first need to understand the difference between owned and earned media.

Owned media includes blog posts, email newsletters, LinkedIn posts, and every other piece of content within your direct control. It drives awareness, and should give your audiences the clearest picture of what the problem is and how you can solve it. Owned media answers the "What does X company do?" prompts on LLMs.

Here's one data point about owned media that I think should wake a lot of founders up:

1 in 10 AI responses directly cites LinkedIn content. That makes LinkedIn the second-most-cited domain by AI — more than Wikipedia, more than YouTube, more than every major news outlet. Only Reddit ranks higher. (I'll give you a moment with that one.)

This means what you're posting on LinkedIn is showing up in conversations your potential customers are having with ChatGPT, Gemini, Claude, and Perplexity. Right now. Today. For queries you'll never see.

And the bar for this is lower than you think. The average cited post has just 15 to 25 reactions. You don't need to go viral. You just need to be consistent and findable.

Earned media is press mentions, UGC, and those random conversations on Reddit — all content that gets published without your explicit approval. It drives authority, and signals to your audiences, including LLMs, that your company is worth considering for insert your value proposition here. When someone asks "What companies are the best for XYZ?" earned media is what gets your name on the list.

With traditional SEO, you were optimizing your own pages. Stuffing keywords on your site. With AI search, the game is predominantly off-page. You're optimizing for what other people say about you.

Brand mentions are the new backlinks.

That's a fundamental shift that cannot be understated. Because it means the best thing you can possibly do to earn the attention of LLMs is get other people talking about you.

What this means for the content you're already creating

There's no such thing as throwaway content anymore.

The "About" section of your website better line up with what you told that reporter about your company last week. If an influencer mentions how much they love a specific feature of your product, it might be time to feature it more prominently, everywhere.

Here are the principles that are emerging for how AI decides who to trust and cite:

  • Repetition across credible outlets matters more than one big hit. A dozen solid brand mentions in trade publications will beat one flashy feature in a tier 1 national.
  • Third-party language carries more weight than your own website copy. What others say about you is more influential than what you say about yourself.
  • Vague positioning trains vague outputs. If you're fuzzy about what your company does, the AI will be too.
  • AI rewards facts, not claims. "We're the leading platform for X" means nothing to a model. "We've processed 10 million transactions" is something it can work with.
  • Consistency beats cleverness when machines are reading. Every interview, every press release, every LinkedIn post is contributing to a body of evidence about who you are. Don't contradict yourself, and the LLMs won't either.
  • If you don't define your story, the model will. And it will pull from whatever it can find — which may not be what you want.

The story you build now is the one AI will keep telling later.

A quick test you can run today

Maybe you've already been playing around in this sandbox, or maybe you're wondering if you're visible at all. Either way, there's a quick way to find out where you stand right now with LLMs, and set a baseline for any work you do to improve their impression of your company moving forward. Here's the playbook:

  1. Create 50 sample prompts where you'd plausibly want to show up in the AI responses. (The shortcut: go to ChatGPT or Claude and say, "Imagine you are a marketer for <insert your company here> — what are 50 prompts you'd want to show up in an LLM for?" Let it do the work.)
  2. Track your baseline. Use an AI search analytics tool to see how you're currently appearing for those prompts. There's a number of them available — Profound, Scrunch, or check out BrandVisibility.ai from a fellow startup founder I know :)
  3. Publish a piece of content. Ideally, this is a piece of earned media, but (as explained above), this is harder to control. If you have been planning a content revamp, or have a new blog you're really excited about, make sure you do steps 1 and 2 before you hit publish.
  4. Track the same prompts again. Where are you now showing up that you weren't before?
  5. Package it for stakeholders. "Because of this piece of content, our brand is now showing up for these queries in ChatGPT/Gemini." That's a story you can tell — to investors to help raise for a PR push, to employees to drive motivation for more content, or even just to yourself the next time you think about pushing LinkedIn down the priority list.

Most companies can't yet attribute real new business directly from LLMs. But every leadership team knows this is where search is heading. Showing that you're thinking ahead — and can prove small wins — earns you credibility you didn't know you needed.

So when I'm working on a press release now...

I ask a better question. Not just "who is this for?" but "how will this be perceived?" Not just by humans, but by the systems shaping how humans find you.

It's more important than ever to be intentional about what you're putting into the world. Because now it's not all just being read. It's being remembered and referenced in conversations you're not a part of.

The companies that understand this early won't just show up in search results. They'll shape the answers.