Human-first PR: why authentic storytelling outperforms AI-generated content

Ask ChatGPT, Gemini or Google’s AI Overviews to name the best builders, hairdressers or even marketing agencies and you’ll get a confident shortlist within seconds. The brands on that list have something in common, and it usually isn’t the biggest marketing budget. Instead, it’s more likely that they’ve been written about, quoted, and referenced by publications that the AI models were built to trust.

Coverage used to be about reaching readers; now it also feeds the systems that answer questions before those same readers needs to make a click or visit a website. This blog looks at why brands that earn genuine editorial coverage are dominating AI answers, and why using AI to handle your press releases tends to backfire harder than many teams expect.

Why this matters right now

AI answers are absorbing the clicks that used to flow to websites. The Pew Research Centre found in 2025 that when Google showed an AI summary, users only clicked through to a traditional result about 8% of the time, roughly half the rate seen when no summary appeared. So, for every 100 searchers you might be showing for position 1, and have a killer site title and description built for high click-through-rates, you’ll only get 8 clicks. And that’s if the keyword isn’t in demand enough to have a sponsored section, visibility through SERPs is shrinking.

Then from the other side of the SERPs, Reuters Institute’s 2026 trends report, expects that news publishers expect search referrals to fall by more than 40% over the next three years. That’s fewer clicks and less traffic coming through organic search for your brand.

The question marketers should be asking has shifted. It’s less “how do we rank?” and more “when an AI builds an answer about our category, are we named in it?”. That answer is decided not just by what your website and content say about your brand, but also by what other people have published about you, which puts PR back at the centre of the visibility conversation.

The brands AI talks about are the brands publications talk about

Let’s get to the basics of how an AI tool builds a response. When dealing with a prompt, it draws on sources it already considers credible, then names a handful of them. Muck Rack and Generative Pulse looked at exactly what those sources are in their 2025 “What Is AI Reading?” analysis. Their report found that more than 95% of the links cited in AI answers were unpaid, around 85% of those were earned media, and close to half of all AI responses included at least one earned-media citation.

The material these systems lean on to describe brands is editorial coverage, not advertising and not what you’re writing on your own homepage. If a credible journalist has written about you, or a credible influencer has talked about you, an AI is far more likely to repeat it than anything else.

Volume of coverage seems to make an impact as well. An Ahrefs study looked at 75,000 brands in 2025 and found that branded mentions across the web positively correlated with AI Overview visibility at 0.664, a strong relationship for this kind of data. The more often trusted sites reference your brand, the more often you surface when an AI fields a question about your sector.

Spreading that coverage across titles compounds the effect. Some research has even found that distributing a story across a wide range of publications could lift AI citations by as much as 325% compared with publishing it on your own site alone.

One placement from a publication which isn’t on your website is useful. Coverage across a range of publications your audience and the AI both trust is far more powerful. Repeat it multiple times, and that’s your answer to how to get mentioned in AI. It’s the work that sits at the heart of a strong digital pr service that knows what it’s doing.

It’s valuable because it’s hard to do

We’ve just made a very clear point that employing Digital PR activity is your main route to getting cited in AI. But the increasing problem for a modern marketer is lack of time. AI has opened the door to solving this problem. A tool can draft a release in seconds, but the problem shows up at the other end, in the inbox of the journalist you actually need. One of the biggest parts of Digital PR are the relationships, if a journalist sees you as someone who sends obvious AI generated press releases and comments, they will lose trust in what you have to say and potentially ignore you in the future.

In this next section, we’re going to put you off using AI for press releases, especially if you need it done right.

Journalists can spot it, and it changes how they treat you

Medianet recently released their 2026 Media Landscape Report, which surveyed more than 800 journalists. 78% said receiving AI-generated pitches lowered their trust in the PR content they were sent, and 48% believed they could almost always tell when a pitch had been written by AI. Close to half of reporters in the study dismissed machine-written releases as lazy and untrustworthy.

That reaction carries more weight than it used to, and that’s because press releases have become the single most-used story source for journalists. In the same report it was named by 86% of them. It gives brands an edge to getting into the press as it’s the format reporters rely on most, but they’re quickest to bin when it reads like a machine produced it.

You don’t sound unique

AI tools are built to generate the most probable phrasing, you might have heard it referred to as “fancy predictive text” or “spicy autocomplete”. Which means that everyone using them lands in roughly the same place. Your release starts to sound like your competitor’s, and theirs sounds like yours this is because by its very design, LLMs are creating average content.

Newsrooms have noticed this flood because they read and write for a living, they can spot the patterns. And yes, these patterns change but the time you spend on making a piece of content not sound like AI, would you have been quicker doing it yourself in the first place?

You strip out the angle that earns the link

Coverage gets earned when you offer something a reporter can’t get anywhere else – aka, proprietary data. Genuine expert viewpoints, a story with a real person at the centre.

AI is good at rephrasing what already exists and poor at originating any of that. Lean on it to write the whole release and you tend to remove the very thing that made the story worth covering, and worth citing months later.

What human-first PR looks like in practice

Okay that sounded pretty anti-AI but we wouldn’t be a sane digital marketing agency if that were the full story. None of this means switching AI off completely. But it does mean using it where it’s strong and keeping people in charge of the parts that build genuine trust.

Put AI to work on the grunt stuff: summarising research, building media lists, proofing copy, spotting trends early. Keep humans on the angle, the interpretation of the data, and the relationship with the journalist.

When you pitch, make it specific. The best way to build trust with a journalist you haven’t worked with before is to show you understand what they actually cover and you can make their workload lighter and readers happier. A tailored human pitch beats a polished generic one every time.

Build your PR around assets worth citing like original research, a data-led story, a named expert with a real point of view, a genuine customer outcome. These are what journalists link to and what AI systems pull into answers later.

Pair that earned coverage with strong content and SEO services on your own site, so the story holds up wherever an AI eventually finds it.

Next steps

AI is changing where visibility is won, but not what wins it.

The brands showing up in AI answers are the brands real publications chose to write about, and those choices are still made by people who can smell a shortcut from the subject line.

Run a quick test this week: ask ChatGPT, Gemini and Google’s AI Overviews to recommend the best brands in your category and see who gets named. If it isn’t you, the answer is coverage worth earning, and that still starts with a human story a journalist genuinely wants to tell.