Simon Murphy • January 22, 2026
Six Rules for PR in the AI Search Era (An Antidote for Chaos)

Just before Christmas, Arun Sudhaman wrote "Spectres at the Feast," unpacking Sir Martin Sorrell's BBC Today programme commentary that "there's no such thing as PR anymore." Sudhaman's diagnosis cut deeper than Sorrell's provocation: the real problem isn't whether PR is dead, but whether the industry can reclaim authorship of its own mandate while better-resourced marketing players claim "scaled, digital storytelling" as their territory.


If you work in communications, this should have landed with more of a thud than the usual industry chatter. The strategic high ground of earned storytelling, surely the very thing PR was built to own, is being contested while we're still explaining what we do.


Rather than fuel the "dead or alive" argument, I've decided to treat this moment as a prompt. After 25+ years watching this industry evolve from "press release and prayer" (there is truth in this) to something far more sophisticated and multifaceted (which you and I both know it is), I keep returning to a simple future looking question: how do we build earned credibility that survives in a world where machines increasingly do the summarising?


Seeking inspiration over Christmas, I re-read Jordan B. Peterson's 12 Rules for Life (2018), which argues for choosing order over chaos by starting with what you can control. Six of his rules struck me as unexpectedly helpful signposts for communications in response to this question.


1) Stand up straight with your shoulders back

PR has a habit of underselling itself, then wondering why it gets treated as a cost centre.


If we want a seat at the table, we need to show up like we belong there. That means describing our work in terms that matter to the business: risk reduced, regulatory heat lowered, reputation enhanced, customer confidence sustained, purchase decisions accelerated.


The final outcome is especially salient. As the AI search era gathers pace, earned credibility is a commercial driver. Recent research from Onclusive found that brands with highly positive earned media sentiment are three times more likely to win in head-to-head commercial comparisons. When your organisation is being described by AI answer engines using other people's words, the quality of those third-party words matters more than ever.


Show your stakeholders what your organisation's ChatGPT profile looks like versus your competitor's. That gap becomes your business case.


2) Set your house in perfect order before you criticise the world

This rule is painfully practical. Before complaining about Sorrell's barbs, let's fix the basics.


Most organisations still have trust gaps hiding in plain sight: dated FAQs, vague positioning, thin "about" pages, leadership profiles that say nothing, verbose thought leadership that says even less. Getting your house in order - namely your own content - isn't just good brand housekeeping. It's the very foundation that feeds the machines.


Generative engines reward clarity, consistency, and credible sources. If you want to show up well in AI-mediated discovery, make your backyard legible:


  • A well-structured website explaining what you do, who you serve, and what proof you have
  • FAQs answering real stakeholder questions, not the ones you wish they asked
  • Clear, quotable narratives for leaders, with verifiable supporting facts
  • Useful, authoritative content that earns its place rather than fills space


Research from my UK based partner Hard Numbers shows that large language models source information from a blend of editorial media, owned content, and other high-authority online sources. This makes Gen AI a force multiplier for existing content efforts. It's not glamorous, but it's compound interest.


3) Tell the truth, or at least don't lie

Trust is significantly undervalued in today's climate, which makes it worth the investment.


In an AI-shaped information environment, credibility is currency and contradictions are tax. Applied habits of yesteryear - massaging reality, spinning weak stories, hiding behind euphemisms - don't just look bad anymore. They're expensive own goals.


The strongest communications teams aren't the ones who "win the narrative." They're the ones who help organisations behave in ways that can be defended, explained, and repeated without regret to both human and machine stakeholders, all the while building trust.


Simple test: would we be comfortable seeing a questionable claim quoted back to us on a chatbot, out of context, in six months? If not, tighten it or drop it.


4) Be precise in your speech

Vagueness can be the more comfortable path, but precision is about being understood. It's the difference between "we are committed to innovation" and "here's what we changed, why we changed it, and how we measure whether it worked."


In generative search, precision is also how you become citeable. If AI systems increasingly act as the first filter between stakeholders and your brand, then "machine-readable credibility" becomes part of the job. Fewer fluffy adjectives, more concrete proof points, more conversational straight talk that can survive repetition across sources.


The goal isn't volume. The goal is reliable information that scales.


5) Make friends with people who want the best for you

Earned has always been a team sport.


In the old world, this meant fostering relationships with journalists, editors, and producers. That still matters, but the stakeholder circle has widened. Today your credibility is shaped by analysts, academics, creators, industry advocates, government, NGOs, employees, customers, traditional media, and niche communities.

The modern earned play is less "spray and pray" and more "build a citation-rich ecosystem." You want credible third parties saying consistent, positive things about you for reasons that stand up to scrutiny.


Internally, the same rule applies. If PR wants to lead in the AI search era, it can't sit downstream of decisions. Build allies across product, legal, customer, HR, risk, and the C-suite, not to ask permission, but to shape reality before it hits the headlines or the model.


6) Do not bother children when they are skateboarding

If PR is to evolve, we must protect experimentation.


Marketing figured this out with the 70:20:10 rule. As Rory Sutherland notes, you need the 10% of honeybees following random routes to find new pollen when the old supply dries up.


Communications has been too risk-averse to adopt this discipline. We need to fix that.


Most communications teams I know run flat out just to keep up with the cycle of work coming their way, which means change happens only after something breaks. That's a rough way to live.


Fortune favours the bold. Ringfence time for structured tests in AI search optimisation, not gimmicks or prompt hacks, but real experiments that teach you what gets remembered and repeated:


  • Which thought leadership formats earn citations, not just clicks
  • Which proof points travel through third-party sources
  • Which questions stakeholders actually ask AI tools about your category
  • Which owned assets improve clarity and reduce confusion


Small falls are the price of progress. Better to have them in controlled pilots than in crises.


Reclaiming the Mandate

Can the PR sector reclaim authorship of its mandate, demonstrating that its counsel and capabilities are indispensable? That's the right question for 2026, but it requires us to stop asking for permission and start demonstrating competence.


The "press release and a prayer" version of PR is mercifully long gone. The discipline that's emerged - earning trust, building understanding, architecting reputation in an age of algorithmic mediation - is more valuable and measurable than ever. But as the AI search era gathers pace, the strategic high ground of earned storytelling is not ours by default.


We have to take it.


This article first appeared on Earned First on 20 January 2026.