The similarities and the differences between pitching for bots and humans.
Whether you’re trying to catch the eye of a top journalist or a competitive AI search algorithm, original data is a great way to stand out.
According to research from Muck Rack released at the end of 2025, the press releases AI cites have twice as many statistics cited as the average press release.
“Opinion data is exactly the type of information that AI-based search really thrives on and that AI search really craves,” said Nathan Richter, senior partner at Wakefield Research. “AI is very good at indexing data, assigning authority to it, and then reporting that back in a kind of a question-and-answer format. So you submit a query and it provides a — usually — a research-backed and cited answer. Opinion data is perfect for that.”
Here’s how to effectively create and package research in a way that will entice AI search without sacrificing human readability.
The human-machine overlap
The good news is that there’s a tremendous amount of overlap between the kinds of content an LLM can easily process and the kinds of content people like to read.
“I think if you ask editors and producers and … we talk about some of these elements of what makes for good AI communications, they would go, ‘Great. That works for me too, as a human being reporting on things.’ It’s just very serendipitous that this approach also benefits optimizing communications for AI,” Richter said.
These items include:
- Be clear and concise, frontloading content with the most important information. Pay special attention to the first 50-100 words.
- Use formatting that makes it easy for both humans and AI bots to scan, including headings and bullet points. Avoid dense blocks of text – neither bots nor humans enjoy that.
- Recency is king. Both bots and editors want the latest, most up-to-date information. For LLMs, there’s a preference for content published in the last year. For human editors, they’ll want something more recent.
What’s most important to AI
Despite all these similarities, there are a few key points where you’ll want to adjust the research to best appeal to AI.
Because AI search works primarily via questions by a person to a robot, Q&A formats are particularly effective.
“With news-you-can-use, your communications clearly state and answer a question and that’s highly valued by AI,” Richter explained. “In terms of designing the research, we’re going to encourage our clients to think and embrace that news-you-can-use approach in that setting. There’s really no downside to that, because that also appeals to traditional media.”
AI also strongly values credibility. This can come in a variety of forms, including the trust in the institution connected to the release, but also to the description of the methodology.
“There is a lot of opinion data and research out there that is conducted to a less-than-ideal standard,” Richter said, so communicators must make sure that not only is their research conducted with rigor, but it’s also described clearly for the AI to access.
Indeed, properly optimizing data for AI means including far more than you might if the release was being sent to a human reporter or editor. With humans, the goal is traditionally to draw them in with a surprising and clever stat and get them to ask for more information. AI can’t be curious and it can’t ask for more information. It can only go off what you feed it – so feed it well.
“To really optimize for AI, what we recommend is including a lot of that data in your release,” Richter said. “And we can’t, of course, forget that it has to be concise and that we want to make sure we’re using our bullets and directing the AI toward the pertinent information in the release, but you can include a lot more information.”
To keep it concise even as you add information, cut out all that clever wordplay and punchy headlines. AI wants just the facts.
“I encourage everyone in that scenario to think of (these) communications less as a press release and more as an exercise in the efficient organization of information through headings and bolding and bullets,” Richter said.
Allison Carter is editorial director of PR Daily and Ragan.com. Follow her on LinkedIn.
The post How to craft data AI search loves appeared first on PR Daily.
