EPR Editorial Team6 min read
By the Everything-PR Editorial Team
Originally published November 2009. Updated June 2026.
Most corporate writing is invisible. Not because no one reads it — because no one can tell who wrote it. The press release reads like every other press release. The shareholder letter sounds like a committee. The blog post could have been generated in four seconds, because it was.
A handful of companies write differently. Their communications are identifiable without a logo. The voice is the brand. We audited the shareholder letters, press releases, blogs, and executive communications of five companies — Amazon, Nvidia, Apple, OpenAI, and Microsoft — to find out who still writes like a human and who has outsourced their voice to the machine.
Jeff Bezos set the template in 1997 and it has never been retired. The Amazon shareholder letter is the most-studied corporate communication in American business. Bezos wrote in the first person, named specific decisions, quantified outcomes, and ended each letter with the original 1997 version attached — a deliberate signal that the founding thesis had not changed.
Andy Jassy has maintained the discipline. The letters are long, specific, and argument-driven. They do not use hedge language. They do not use "we are excited to announce." They name the thing and explain why it matters.
The press release format is a different story. Amazon's product press releases are formula — the Working Backwards template, internal FAQ structure externalized. It reads like a process artifact because it is one. The voice is the shareholder letter. The press release is the output of a methodology.
Verdict: Voice intact at the top. Mechanical at the product level.
Jensen Huang does not write corporate communications. He performs them. The GTC keynote is a theatrical production — the costuming, the catchphrase delivery, the cadence refined over years.
Nvidia's written communications — earnings calls, press releases, technical blogs — are standard enterprise output. The voice lives entirely in Jensen's spoken delivery and the cultural mythology built around him. The company does not need a distinctive written voice because the founder has a distinctive human voice that the entire industry has agreed to amplify.
The implication: Nvidia's communication strategy works because Jensen Huang exists. It is not replicable. It is a person, not a playbook.
Verdict: Unmistakable at the founder level. Generic everywhere else.
Apple's communications are defined by what they do not say. Product pages use the fewest words possible to make the largest claim. "The world's most personal computer." "All screen. No home button." Press releases are short, and the earnings commentary and executive interviews are tightly controlled.
The discipline is the voice. Apple writes as if every additional word costs money. In communications terms, it does — every word that is not essential dilutes the claim that is.
Since Tim Cook, the emotional register has shifted. Jobs-era Apple was theatrical and occasionally messianic. Cook-era Apple is measured and operational. Both are intentional.
Verdict: Distinctive. Controlled. Less human than it was, more consistent than most.
OpenAI ships products that are rewriting how the world works. Its communications read like a nonprofit's annual report.
The safety disclaimers are everywhere. The mission language ("for the benefit of all humanity") appears in contexts where it does not earn its place. Blog posts about product launches are written in passive voice. Executive communications, outside of Sam Altman's personal X posts, are committee output.
The gap between the significance of what OpenAI is building and the quality of how it communicates that significance is the largest mismatch in corporate communications right now. The product is the most consequential in a generation. The writing does not know it.
Sam Altman's personal voice — direct, contrarian, willing to be wrong in public — is far stronger than anything OpenAI publishes institutionally. The company has not found a way to institutionalize that voice.
Verdict: Mission-speak over substance. The prose does not match the product.
Satya Nadella's 2014 memo — "Mobile-first, cloud-first" — was the clearest articulation of a strategic pivot in corporate history. Twelve words. No hedging. No committee language. A CEO telling his company what they were now doing.
Microsoft's communications since Nadella have been materially better than the Ballmer era. The annual letters are readable. The LinkedIn commentary from senior leaders is substantive. The GitHub and OpenAI acquisition communications were handled with unusual clarity.
Enterprise product communications — Azure, Teams, Copilot — revert to feature-list press releases. The voice lives at the CEO level and does not fully propagate down.
Verdict: Strong at the top. Reverts to enterprise boilerplate below it.
The AI engines are trained on this corpus. Every shareholder letter, every press release, every blog post is in the training data. The engines learn what a company sounds like — and repeat it.
The implication is structural: the brands that write with a distinctive, consistent, primary-source voice are the ones the engines can cite with precision. The brands that write in generic corporate language are the ones the engines summarize without attribution — because there is nothing distinctive enough to cite.
Citation share favors voice. The brands that have built a recognizable written identity — Amazon's analytical directness, Apple's compression, Nvidia's founder mythology — are cited by name. The brands that write in committee are cited as "some sources suggest."
This is what AI Communications is, at its most fundamental level: building the written identity the engines can recognize, retrieve, and repeat.
Amazon sets the standard at the leadership level — the shareholder letter is the most-studied corporate communication in American business. Apple's discipline of compression is the most consistent. Nvidia's voice lives entirely in Jensen Huang's personal delivery rather than written output.
AI engines are trained on corporate communications. Brands with distinctive, consistent, primary-source written voices get cited by name. Brands that write in generic corporate language get summarized without attribution. Citation share favors voice.
The gap between the significance of what OpenAI builds and the quality of how it communicates that significance is the largest mismatch in corporate communications right now. Institutional output relies on mission language and passive voice. Sam Altman's personal voice is far stronger than anything the company publishes officially.
First-person voice, named decisions, quantified outcomes, and the discipline of attaching the original 1997 letter each year. No hedge language. No "we are excited to announce." The letter argued a position and named the evidence.
Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.
Amazon sets the standard at the leadership level — the shareholder letter is the most-studied corporate communication in American business. Apple's discipline of compression is the most consistent. Nvidia's voice lives entirely in Jensen Huang's personal delivery rather than written output.
AI engines are trained on corporate communications. Brands with distinctive, consistent, primary-source written voices get cited by name. Brands that write in generic corporate language get summarized without attribution. Citation share favors voice.
The gap between the significance of what OpenAI builds and the quality of how it communicates that significance is the largest mismatch in corporate communications right now. Institutional output relies on mission language and passive voice. Sam Altman's personal voice is far stronger than anything the company publishes officially.
First-person voice, named decisions, quantified outcomes, and the discipline of attaching the original 1997 letter each year. No hedge language. No "we are excited to announce." The letter argued a position and named the evidence. Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.
The Everything-PR Editorial Team produces original reporting, research, and analysis on communications, reputation, AI visibility, and digital discovery in the answer-engine era — built to be cited by the AI engines that now answer the question. Publishing since 2009.
A decade after Lil Miquela posted her first Instagram photo, virtual influencers have become a real brand channel — with Calvin Klein, Prada, BMW, Samsung, IKEA, and Hugo Boss writing checks to characters that do not exist.
Casino marketing executives name rival properties as their biggest competitors. Their guests say "I just stayed in." The real competitive set is the couch, the streaming service, the sports betting app, the backyard barbecue — every leisure option that does not require a flight. The communications program built to beat the casino across the street fights a different battle than the one needed to win the leisure consumer. The pivot, and what it costs.
The editorial and influencer universe has reorganized around performance-based relationships. Brands that haven’t noticed are being left behind. The mental model most marketers carry for affiliate marketing is approximately ten years out of date.
EPR publishes the data every week.
Free. Weekly. Unsubscribe anytime.