Last week, I had to do a lot of research on a new client in a hurry because they were in the middle of a pretty substantial PR crisis.
In the past, I would have plugged their name into Google and scoured page after page of search results looking for every piece of information I could find about them. Websites, news, social media activity, reviews, and even posts on forums—I want everything I can get my hands on to get a sense of how they normally conduct themselves, what kind of person they are, and what’s already out there that I could use to combat the crisis.
Doing this research properly the old way can take hours.
But today, it’s a lot more efficient because I can often leverage the power of AI to quickly find and summarize everything that’s out there.
This opens a proverbial Pandora’s Box, though, because now we also have to account for what shows up in these Large Language Models, also known as LLMs. People in the tech space love their acronyms, apparently.
This adds an entirely new facet to public relations.
But while these tools are powerful, they’re still far from perfect. Every LLM in use today has a serious problem with hallucinations, which is a gentle way of saying that they often just make shit up. Even if you don’t regularly use AI, I can promise you’ve seen this first hand. Google’s AIOs, an acronym for AI Overviews, are a perfect example of this. These are the blurbs that show up at the top of the search results to provide a summarized answer to your query, and they’re widely hated by users.
A few examples of the incorrect information I’ve seen in these AIOs includes:
- Gouda cheese is the best-selling cheese in the U.S.
- To stop cheese from sliding off your pizza, you should add ⅛ cup of nontoxic glue.
- An effective cure for kidney stones is to drink urine.
To say these hallucinations are problematic would be an understatement.
While the hallucinations are less severe when it comes to facts about people and companies, they are still common. Case in point: about a year ago, I asked Grok to tell me about myself, and it insisted that I had served in an Army Airborne unit, when in reality, I served in the Marine Corps.
Today, more people are turning to AI tools like ChatGPT, Grok, and Claude to find information they would previously have gotten from a traditional search engine, and that number grows every day. That’s why it’s essential to ensure that both your personal and company brands show up when people ask about them in these tools. It’s also important that they show up in the answers when people ask questions about your industry.
So how do we achieve this?
Well, unfortunately, it’s not quite as simple as prompting the systems with the information you’d like them to include their answers. Let’s be honest—if it were that easy, bot farms would be spun up overnight by shady marketers to spam these systems into oblivion.
The LLMs get their information by crawling the open web, but they don’t just incorporate everything they find. Key data about people and companies comes, in part, from their official websites, but that doesn’t carry anywhere near as much weight as the information that comes from certain trusted third party websites.
Legitimate media outlets are the backbone of this information, making effective public relations even more important than ever, and there are two key components to this. This first and most important is editorial features that share your unique insight about your industry, and the second is press releases.
Editorial features are the most important because they are less susceptible to manipulation. Real publicity is difficult to earn. You have to start with a topic that impacts a lot of people, find a unique angle to pitch, and then develop and nurture relationships with journalists to turn it into an engaging story. That’s tough to game, and it’s exactly why it plays such an outsized role in getting your brand to show up in the responses these LLMs provide to users.
Generally, the more information you’re able to include in these editorial features, the better. This is both because it gives the LLMs more information to work with, and because it demonstrates your importance.
Think about it like this—if a journalist only includes a sentence or two from you, that says something about the impact they feel your insight has for their audience, but if a majority of the article is made up of your thoughts, clearly they feel your insight is far more valuable. This is also why it’s so important to not only get quoted in editorial features, but also to seek out comprehensive interviews and contributor opportunities as well. The latter gives you the ability to share more of your unique insight.
Press releases also play a role, albeit, a much smaller one.
That’s because while they are a paid channel that anyone can be cited in, there is still a degree of editorial oversight. Typically, the editors at the wire services that publish these press releases will verify that the claims made in a press release are free of hyperbole, followed by basic fact checking, but it won’t usually be as comprehensive as what you’d expect from a journalist. Once they’ve done their due diligence, the press release will be published and then syndicated to partner media outlets.
The key here is to ensure any press releases you submit are, in fact, newsworthy topics. It’s not enough just to put the information out there. It needs to be useful and engaging. Plus, the more effective your press release is, the more likely it is to generate additional editorial features from journalists looking for topics to write about.
It’s important to note that while LLMs do get their information from a variety of media formats, text based articles currently produce the most impact, both in terms of the information itself and how quickly it makes its way into the systems.
AI has radically transformed public relations, but the foundation is the same as it’s ever been. We still have to work to earn high quality editorial features just like we did before AI. The main difference today is that those features are no longer the end goal. Today, the end goal is to demonstrate authority status so that we also show up in the AI platforms that a growing number of users rely on.
Those who take action now will gain a massive competitive advantage, while those who don’t will quickly fade away into obscurity.
