The fundamentals of comms still apply: know your audience, articulate a story, speak in plain language, answer specific questions
LLMs are becoming the new “front page.” In the same way Google replaced the news, LLMs seem to be doing that to search.
LLM stands for large language model and it’s effectively the infrastructure that enables generative artificial intelligence (genAI) to work. As such, getting “placed” in an LLM may well be an emerging task for public relations professionals and communicators.
And that conjures this important question: How can communicators pitch an LLM?
I submitted a query through Peter Shankman’s new Source of Sources service and asked that very question. Below is a bit of what I heard from professionals who have been experimenting with pitching genAI.
1. Articulate a unique point of view
Notified CMO Adam Christensen indicated pitching an LLM comes down to the basic skills communicators already possess: Crafting a unique point of view and communicating that view in a way that appeals to the intended audience.
Having a story is the “most important” aspect he said over a Zoom call. “Can you articulate that story?” he asked rhetorically. “Can you place that story and then optimize the rest?”
Christensen, who started his career in the communications trenches, said press releases are currently the “fastest way to get content” into an LLM. This is because “LLMs are prioritizing the news” and in a sense, a press release is an official, and ostensibly sanctioned, public statement put forth by a company.
“I think corporate news, especially when it’s an official statement by a brand, continues to be able to one of the bigger sources [of exposure] for customers,” he said.
Notified doesn’t have “empirical data” to prove LLMs favor press releases. However, Christensen says the company has seen this anecdotally. Further, Notified’s conversations with genAI vendors seem to validate his hypothesis.
It’s not a stretch from where I sit, either.
A large press release distribution platform is a reasonable source of data on which to train LLMs. Relative to the average quality across the internet, the writing is probably better. Further, press release distribution vendors do provide some level of vetting before publishing a release, though it’s far from foolproof.
Notified owns GlobeNewswire, which is among the larger press release distribution providers. In my experience, Globe has better distribution in EMEA than the other large rivals. I haven’t experimented with the service in the U.S.
2. Answer very specific questions
Marcus Sheridan once turned around a flailing small business with one simple strategy: answering customer questions on a blog in plain language. Since search is an explicit expression of need – just like a question – Google loved his content. I think something similar is happening with genAI.
Sean O’Meara, the Managing Director of Essential Content in the UK, might see that too. He said his company has had some success, too. Google’s AI overviews, for example, have included his firm’s ideas (the screenshot nearby is his).
“Our approach was simply to analyze the content formats that Google’s generative AI prefers and match query type to content formats,” he said in an email. “We then identified five core characteristics of content that the AI liked.”
Those five core elements are:
- Use specific questions as headlines and provide specific answers;
- Develop informative visuals – like comparative graphics;
- Put strongly formatted content summaries front and center;
- Cite your source material and make it easy to identify those sources; and
- Be specific – specific content outperforms the broad content of his competitors.
If this sounds similar to SEO, I think that’s probably true with one big difference: generative AI is like semantic search – which aims to understand meaning based on context – only it does it much better than traditional search engines.
(click image for higher resolution)
3. Basic comms principles, including traditional media, still matter
“AI search engines are trained on high-authority media sites, making press coverage more important than ever,” according to Andrew Holland, an SEO director for a Manchester, England-based marketing firm called JBH. “The more often you’re mentioned in top-tier publications within your area of expertise, the better your chances of being recognized by AI.”
That’s the earned-media component, but owned media has a role too: “LLMs frequently pull from list-based content when generating recommendations. To ensure your brand appears in AI-generated lists, you should” create your own lists and strive to be listed in credible third-party listicles too.
Traditional SEO may still have a role too, according to Holland. “Many LLMs rely on Bing and Google to retrieve results, meaning traditional SEO still plays a critical role in AI search.”
However, it also offers an end-around Google’s system of search rankings. This is because “AI search engines allow users to bypass Google’s traditional ranking system, disrupting the paid search model.”
“AI search isn’t replacing SEO and PR, it’s making them even more essential. To rank in LLM results, brands need press coverage, authoritative citations, and strong SEO. The more AI sees your brand in credible sources, the more likely it is to surface you in LLM search results.”
If Holland’s ideas sound like traditional integrated marketing, then you are probably right. The fundamentals of good marketing and communications haven’t changed: The priorities of marketing were, and still are, owned, shared, earned, paid – in that order.
* * *
>>> Subscribe by email for free: Check out my weekly blog posts, weekly podcasts, or a monthly newsletter that rounds up interesting reading from the last 30 days (examples).
If you enjoyed this post, you might also like:
Build marketing and PR programs rather than executing campaigns
Image credit: ChatGPT / DALL-E
