The category you rarely benchmarked — but should have

Why PR is a critical input to AI — and a strategic imperative for sourcing teams

Will PR be the first marketing category to be replaced by AI?” It’s a fair question — especially given how easily AI generates press releases and blog posts. For many, PR felt ripe for automation: word-heavy, reactive, and often invisible in ROI frameworks (do not blame the messenger here).

In addition to its position within the marketing mix, PR has, in recent years, often taken a bit of a beating on peer networks such as LinkedIn and other social media networks. Stories from PR executives, agency staff and others range from accounts of “how PR is losing its perceived value to marketers”, “how agencies are seeing plunging margins” and “how brands are pushing scope of work pricing towards the bottom” has regularly come across my feeds for some time now.

What used to be output is now input

Historically, PR has often been a reactive lever in the marketing mix. It became a tool when companies needed to contain bad press, clean up a leadership misstep, or respond to sudden scrutiny. For many procurement leads, PR was a category with low visibility, softer metrics, and often vague deliverables. You wouldn’t find it at the heart of long-range commercial planning. On top of that, it is usually not a very large monetary spend category.

That’s no longer true. In an AI-first world — where visibility is shaped not by your website or paid ads but by what credible sources say about you — PR is now quickly becoming the connective tissue between your brand and the algorithms that summarize it. For everyone, this is a fundamental shift in how influence is built and measured.

PR looked replaceable — until it wasn’t: AI excels at replication but lacks cultural awareness, timing, and judgment. It can’t grasp why a regional misstep triggers outrage in one market and shrugs in another. PR is about escalation, timing, and intuition — elements machines can’t mimic.

From SEO to LLMs — why PR now owns the moment

For the last 15 years, SEO ruled the discovery game. Get your keywords right, build backlinks, optimize page speed — and you had a shot at ranking on page one. PR and SEO teams ran in parallel, occasionally aligned, rarely integrated.

Now, enter large language models (LLMs) — the foundation of tools like ChatGPT, Perplexity, Claude, and Gemini. Unlike traditional search engines, these models don’t crawl the web in real time. They synthesize answers from static training datasets built months or even years ago. And what’s inside those datasets? Not your latest web update or a fleeting social post. It’s trusted, high-authority content — including news stories, analyst commentary, and thought leadership.

According to OpenAI and Anthropic disclosures, LLMs are trained on massive corpora that prioritize high-quality public sources. A 2024 Stanford study found that over 70% of factual citations generated by LLMs trace back to journalistic or editorial sources, particularly those ranked high in Google’s “E-A-T” framework: Expertise, Authority, and Trustworthiness.

In all honesty, tools like Perplexity, ChatGPT (with browsing), and Gemini offer the closest experience to live web access. However, most enterprise and open-source LLMs do not crawl the web themselves. They rely on what they were trained on — making PR and trusted earned media critical input for brand representation in those models.

This means that the stories PR teams place — especially in mainstream and trade media — don’t just influence people — they shape the worldview of the very machines now acting as consumer and B2B filters.

So when someone asks, “which brands lead in sustainable luxury?” or “who’s setting the tone in AI ethics?”, the LLM isn’t quoting your press release or homepage. It’s summarizing the coverage PR secured — often months earlier. That makes PR the new frontline for how brands are interpreted by machines, and by extension, by decision-makers who trust those machines to summarize the market.

This shift from search-based discovery to answer-based engagement has turned PR from a credibility tool into a discoverability engine. It matters less what your brand says about itself and more what credible third parties say on your behalf. And in an environment where generative tools increasingly power product search, investment research and influencer briefs, earned media has become the algorithmic memory layer of your marketing stack.

It’s not just stickier than paid, or more credible than owned. It’s the content most likely to be cited — by humans and machines alike.

Today’s media environment is fragmented. TV still commands reach, but recall is fading. Influencers provide immediacy, but their content often disappears in a 24-hour cycle. Paid media has scale, but lacks permanence.

PR, on the other hand, creates content that lives on in trusted sources. It forms the citations and summaries that AI systems prioritize. It’s the stickiest part of the media mix.

This doesn’t make PR more important than other disciplines — but it does make it more enduring. And in a machine-readable world, that memory is your most scalable asset.

What this means for procurement and strategy

If you’re managing marketing budgets, now is the time to rethink how PR is scoped, measured, and invested in. For years, procurement leaders have treated PR as a soft-cost category — low monetary spend (compared to other marketing vehicles), high variability, often lacking standard metrics.

But in 2025, PR is one of the only marketing functions that directly shapes your brand’s presence in AI discovery platforms. Even if it represents just 2–3% of your spend, it may account for 10x the influence when it comes to how brands are understood by the algorithms powering modern commerce.

This shift suggests several actions for procurement. Here we list a few:

1. Review your current MSA agreements to make sure you have a modern contract with strong clauses around reconciliation, auditing, governance, third-party expenses and certainly an updated AI policy.
2. Evaluate how legal terms such as insurance, liability, IP etc. are impacted by this shift of focus of PR services.
3. Tighten and update PR scopes of work (SOW) to ensure the right deliverables are included in any scopes.
4. Update SOWs with measurable AI visibility outcomes that matter in the “new world”
5. Include citation tracking and earned-media indexing in agency KPIs
6. Challenge your current PR agencies to drive innovation in this space for your brand. They are the experts.
7. Find and engage agency partners who understand the LLM landscape and its implications
8. Rethink value — not just in terms of impressions, but in terms of algorithmic recall

Traditional PR metrics still have value — but they don’t tell the full story in an AI-first environment. We need to evolve from counting coverage to measuring influence on machine outputs.

When speaking to industry experts, they list examples of how Ai-first KPIs can evolve:

1. Share of Voice → Share of Trusted Mentions in LLM Data Sets
2. AVE (advertising value equivalent)→ Brand Presence in GenAI Recommendations
3. Total Hits → Indexed Coverage Across High-EAT Sources (Expertise, Authority, Trustworthiness)

One Fortune 100 brand recently told us they began evaluating PR agency performance based on generative recall. That is, how often the brand appeared accurately and favorably in AI summaries across platforms like ChatGPT, Perplexity, and Gemini. It’s early days, but the trend is clear: what used to be PR “outputs” are now AI “inputs”.

PR isn’t just surviving the AI transition — it’s gaining huge relevance. Not because it changed, but because the world around it did. The brands that succeed in the next wave won’t be the ones who shout the loudest — they’ll be the ones most remembered by the machines we now trust to filter truth.

That starts with PR. And it requires marketers and procurement leaders alike to see it not as a tactical function, but as strategic infrastructure. Because in a world of algorithmic memory, earned media is everything. And as always — the early bird catches the worm.