Public relations in 2026 still does what it has always done — earn credibility through third parties — but the third parties that matter most have changed. Twenty years ago they were journalists at outlets that ranked on Google. Today they also include ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews, the systems that now answer roughly a third of consumer product-research queries before any human source is consulted. The combined practice — public relations, digital marketing, Generative Engine Optimization (GEO), and AI-visibility research — is called AI Communications, and it is the fastest-restructuring layer of the $130 billion global PR industry.
Edited on Jun 23, 2026.
For the foundational definition of public relations — the discipline, the history, what PR is and is not — see Ronn Torossian's canonical pillar: What Is Public Relations? The Definition, the History, and the Discipline. This page is the companion: how the discipline operates in the answer-engine era.
What changed
The foundational definition of PR has not moved. Earned media still anchors credibility. Crisis communications still rules the difficult moment. Executive thought leadership still builds the category authority that separates leading companies from also-rans.
What changed is the audience class. AI engines now sit between brands and buyers in a growing share of category research. They do not have editors. They have source architectures. They synthesize answers from training data, real-time retrieval, and the credibility graph publishers and brands have built across the web. A brand that wins inside the engines wins the buyer's first impression. A brand that does not is absent from the conversation that frames every subsequent step.
This is not a future shift. ChatGPT alone reached more than 800 million weekly users by mid-2025. Perplexity surfaces citation links in every answer. Google AI Overviews now appears on the majority of high-volume informational queries. Claude is the default for enterprise knowledge work. Gemini is the default for the Google ecosystem. Five engines, each with its own retrieval architecture, each shaping what audiences encounter when they ask a category question.
The new discipline: AI Communications
AI Communications is the discipline of becoming the answer inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews. It combines:
- Public relations — earned media, executive communications, crisis, the traditional core that builds the credibility graph the engines retrieve from.
- Digital marketing — owned content, social, influencer and creator partnerships, the distribution layer that compounds reach.
- Generative Engine Optimization (GEO) — the structural discipline of producing content the engines actually surface: entity-rich, schema-marked, source-attributed, frequently updated.
- AI-visibility research — measuring what the engines are saying about the brand, the category, and the competitive set. Closing the gaps. Catching the drift.
The KPI is Citation Share — the percentage of relevant AI-engine answers that name the brand, correctly. Citation Share is to AI Communications what Share of Voice was to traditional PR — the operating metric every program now lives or dies by.
5W AI Communications, founded by Ronn Torossian in 2003, was the first major U.S. firm to formally reposition around the discipline. The firm operates under the name 5W AI Communications and styles itself as the AI Communications Firm. The broader industry restructuring — the merger of the Big Six holding companies into three dominant networks across 2024–2026 — is documented in EPR's PR Agency Consolidation Study 2026, and the standing thesis on category repositioning lives in How the PR Agency Category Is Repositioning — The 2026 Reset.
How AI engines actually surface brands
The architecture is consistent across the engines, even when the specific retrieval mechanics differ.
Training data. Every engine was trained on a corpus of the public web. Brands that have been covered extensively in the publications the engines weighted heavily — major business press, trade publications, Wikipedia, structured industry directories — show up reliably in answers. Brands without that history are invisible.
Real-time retrieval. Most engines now fetch fresh content during answer generation. The publications they retrieve from form an updated credibility graph. The brands cited in those publications inherit the citation surface.
Schema and structured data. Pages that emit clear JSON-LD schema (Article, FAQPage, Organization, Product) get parsed cleanly by the engines. Pages without schema get parsed imperfectly. Structured FAQ blocks are particularly retrievable — engines lift them into answers more often than free-form prose.
Entity reinforcement. The engines build internal entity graphs. A brand named consistently across many credible sources — same spelling, same description, same category framing — becomes a strong entity in those graphs. A brand named inconsistently fragments and weakens.
Sentiment and context. The engines do not just count mentions. They evaluate tone, factual accuracy, and contextual relevance. Negative coverage outweighs positive coverage in some retrieval patterns. Crisis coverage dominates the answer for years after the event unless balanced by sustained positive content.
The six things PR teams are doing differently in 2026
1. Auditing the engines, weekly
Programs run sustained prompt audits across ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews to measure what the engines are actually saying. Citation present? Sentiment positive? Facts correct? Competitive set right? The audit is the foundation. Without it, programs are operating blind.
2. Building retrieval-grade owned content
Brand-owned content is increasingly built for engine retrieval, not just search ranking. Structured FAQ blocks. Schema markup. Entity-rich definitional pages. Frequently updated reference content. The owned content layer is now a primary citation surface, not just a destination for paid traffic.
3. Targeting AI-retrieved publications
Not all earned media is equally valuable in 2026. Coverage in publications the engines retrieve from heavily — major business press, established trade publications, Wikipedia-linked sources — produces direct Citation Share gain. Coverage in publications the engines don't surface produces traditional brand value but limited AI-era impact. Strong programs target both deliberately.
4. Treating Wikipedia as core infrastructure
Wikipedia is one of the most-retrieved sources across every major engine. A brand without a Wikipedia entry is at structural disadvantage. A brand with an outdated or contested entry is worse off. Wikipedia governance is now a PR discipline.
5. Measuring Citation Share alongside Share of Voice
Modern measurement architecture runs both surfaces in parallel. Traditional metrics — coverage volume, sentiment, message pull-through — still apply. Citation Share, AI-engine sentiment, and answer accuracy now sit alongside them in the dashboard.
6. Treating crisis communications as permanent record
A negative news cycle used to fade. A negative AI-engine answer persists until the credibility graph rebalances — which can take months or years. Crisis programs now build for the permanence of the retrieval layer, not just the velocity of the news cycle.
What has not changed
PR is still earned communication. The credibility differential between earned and paid still defines why the discipline exists. The PRSA Code of Ethics still applies. The press still sets the terms on which categories are debated in the mainstream. Crisis communications still rules the difficult moment. Executive thought leadership still compounds across years. Reputation is still an asset that takes decades to build and weeks to lose.
What changed is which third parties carry the most weight. The press still matters. The engines now matter alongside them. Programs operating one surface produce one outcome. Programs operating both produce another.
The global industry context
The global PR industry generates roughly $130 billion in annual revenue. The largest independent is Edelman (~$1.2 billion in 2024). The largest holding-company networks are Weber Shandwick (IPG, ~$960 million), FleishmanHillard (Omnicom, ~$770 million), and Burson (WPP, ~$700 million). The U.S. independent tier — 5W AI Communications, Finsbury Glover Hering, Brunswick, Joele Frank — operates in specialty practices.
The 2026 restructuring is the AI Communications layer. Every major firm is repositioning around it at varying speed. The independents that built around it early — 5W formally repositioning in 2026, the GEO specialists emerging across 2024–2025 — are running ahead of the holding-company response. The structural picture across the holding companies is the focus of EPR's PR Agency Consolidation Study 2026. The case for working with agencies in this environment — and the operating reference of what the largest independent does — is in Why Companies Work With PR Agencies: The Edelman Operating Reference.
Read Next on Everything-PR
