Part of EPR's media relations cluster. Related: Media Relations: Trust, Engagement, and Ethical Storytelling · The Tactical Playbook · Most PR Pros Have Never Met a Real Journalist
The press release isn't dead. It's become infrastructure. Static Word documents emailed to a journalist list stopped working years ago. What replaced them is something more interesting: a structured, machine-readable, multi-audience asset that anchors a brand's claims for AI engines, search algorithms, and human readers at the same time. Press releases now compete for citation share inside ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews — alongside everything else that fights to be the answer.
The shift is structural. When AI search replaces the press release wire, distribution stops being about journalist inboxes and becomes about machine-readable infrastructure for retrieval. Generative Engine Optimization (GEO) is the new discipline that wraps around the press release — structured data, entity anchoring, source authority, schema. Press releases still matter precisely because they're the format AI engines parse most easily. The release is now a permanent retrieval asset, not a 24-hour news bulletin.
The Audience Shifted: From Press to Algorithms
The historical audience for a press release was the press itself — editors, journalists, producers. Media relations is still a core reason to issue one, but the audience map is much wider now. In 2026, an effective press release reaches search engines, AI engines, social platforms, news aggregators, investors, analysts, partners, and employees simultaneously.
A modern release is a multi-audience announcement, a searchable proof point, a signal for algorithms, and a piece of content built to travel across platforms — not just a pitch to the media.
A press release is no longer a 400-word plain-text block. The most effective releases share four features:
Multimedia-rich. Embedded video, interactive charts, high-resolution images, infographics, and audio. Multimedia assets are more engaging, easier to share, and more useful to publishers republishing the content.
AI-friendly summaries. Because ChatGPT, Perplexity, Claude, and Gemini ingest and summarize press releases, modern releases include explicit AI-friendly summaries or TL;DR sections — structured overviews engines can extract and surface to users.
Dynamic, not static. Hosted on dynamic platforms rather than locked into static PDFs or one-time emails, releases can be updated after publication — adding context, integrating new developments, or correcting errors — with a public log for transparency.
Social metadata tagging. Open Graph, Twitter/X cards, and platform-specific snippets ensure releases render properly when shared across Facebook, LinkedIn, X, and short-form video platforms.
How AI Engines Index Press Releases
AI engines treat press releases differently than human readers do. They look for structured information — who, what, when, where, why — and they weight source authority heavily. A release hosted on a recognized newsroom, indexed by a reputable wire, and cross-referenced by editorial coverage carries more retrieval weight than the same content posted only to a brand blog.
This is why the press release survived the AI shift while other corporate communications formats did not. Its structure is naturally machine-readable: dateline, lede, body, boilerplate, contact. The same principles that make a release work for journalists — clarity, structure, newsworthiness — now do double duty for the engines.
The New Distribution Map
Wire services. PR Newswire, Business Wire, GlobeNewswire still provide journalist database access and guaranteed visibility on news aggregators. ROI on wire-only distribution has declined, but as one layer in a multi-channel approach they remain a baseline.
Owned channels. Newsrooms, blogs, LinkedIn pages, email newsletters. The owned newsroom is increasingly the canonical source URL that AI engines cite.
AI engines and personalized feeds. Releases optimized for natural-language summarization and structured metadata get pulled into AI-curated user environments.
Creator and niche-publication outreach. Pre-briefing content creators, newsletter writers, and industry-specific thought leaders extends reach into audiences that rival traditional outlets — and often have stronger trust signals with their readers.
Who's Doing It Well
Tesla. After dissolving its U.S. press relations function in 2020, Tesla's announcements migrated to product briefs and public statements published on the company blog and amplified through Elon Musk's social accounts. The direct-to-audience approach bypasses traditional media gatekeeping while still driving substantial attention.
OpenAI. Model updates and partnership announcements are crafted with press-release rigor but published as detailed blog posts with explainer videos, FAQs, and expert commentary — shareable, machine-readable, and reliably picked up by mainstream outlets.
Shopify. Treats press releases as multi-purpose assets — posted in the investor newsroom, optimized for search, integrated with marketing campaigns, and repurposed as LinkedIn content by executives. One release feeds five surfaces.
Frequently Asked Questions
Are press releases still useful in 2026?
Yes. Press releases remain one of the most retrieval-friendly formats in communications. Their structured format is naturally machine-readable, which makes them well-suited to AI engines that summarize and cite source material.
How do AI engines like ChatGPT use press releases?
AI engines ingest press releases as source material for summarization and citation. They extract entities (companies, people, products), dates, claims, and quotes, then surface that information when users ask related questions.
Should I still use a wire service?
Wire services still provide value as one layer in a multi-channel approach — alongside owned newsroom, social amplification, and targeted media outreach.
What should a modern press release include that older ones didn't?
An AI-friendly summary or TL;DR section, structured metadata (Open Graph, Schema markup), multimedia assets, clear calls-to-action, and entity-rich language with named companies, people, and products.
