Originally published: April 22, 2013 · Updated: June 16, 2026

In April 2013, The Guardian profiled Lewis Dvorkin, then Chief Product Officer of Forbes Media, who had spent three years rebuilding the company around an idea The Guardian's profile labeled "advertising-supported journalism." Dvorkin had launched BrandVoice in 2010 — the Forbes platform that let brands publish content directly into the Forbes content graph, indexed alongside editorial. Forbes.com traffic had doubled in the preceding thirty months. Newsweek had stopped its print run in December 2012. The original 2013 post argued that Dvorkin was the unlikely figure showing the rest of legacy media how to be profitable in an era where everyone was a content publisher, and that PR professionals could borrow the same playbook — produce content that ranks, gets read, and reaches audiences directly.

Thirteen years on, Dvorkin's bet is the architectural ancestor of every major AI Communications mechanic in market today. The lesson scaled larger than even The Guardian's profile imagined.

What Dvorkin actually built

BrandVoice and its successor AdVoice gave Forbes' largest advertisers — Honda, Capital One, IBM, Microsoft, SAP, Oracle, Deloitte, and others — direct publishing rights inside the Forbes domain. The brand content was clearly marked as sponsored. The editorial content sat alongside it. The Forbes contributor network, which Dvorkin also built, expanded from a handful of staff writers to over 2,000 outside contributors publishing under the Forbes URL. The combined model produced a content surface that grew faster than the editorial budget could have grown on its own and that was structurally optimized for search retrieval.

By the time Dvorkin left Forbes in 2018 to become Chief Content Officer at Tribune Publishing, the architecture he had built was the dominant model for digital publishing. Most major business and consumer outlets ran some version of it. The line between editorial and sponsored content had become a routine product decision rather than a philosophical fight. The journalism establishment did not love it. The economics worked.

What 2013 got right

Three calls held.

Quality content at scale beats editorial scarcity. Dvorkin's bet was that a much larger volume of high-quality content from a broader pool of named experts would outperform a smaller editorial operation. The bet was correct. Forbes' digital audience grew through every subsequent platform shift. The contributor model has been copied or adapted by Inc., Entrepreneur, Fast Company, Adweek, and most major business publications.

Brands could become publishers. The 2013 post argued that PR professionals could use the same playbook to publish content that ranks and reaches audiences directly. That argument turned into the entire content-marketing industry over the next decade. HubSpot built a $20 billion-plus market cap on the thesis. Salesforce, Adobe, Cisco, IBM, Oracle, and SAP all run scaled owned-media operations that produce more reach than their paid media. The earned-media-only PR model lost share to brand publishing every year between 2013 and 2025.

Journalists are not the only authentic verifiers of truth. The original post pulled the quote from Dvorkin and noted that experts bring skills journalists do not. That position was contested in 2013. By 2026 the boundary has effectively dissolved. Substack, Medium, LinkedIn long-form, YouTube essays, and named-expert podcasts now occupy retrieval slots that used to belong exclusively to legacy journalism. The AI engines weight named subject-matter experts alongside named journalists when synthesizing answers.

The 2026 extension: Forbes meets OpenAI

In May 2024, Forbes signed a content licensing agreement with Perplexity AI. The agreement formalized what Dvorkin's BrandVoice architecture had already implied — Forbes content is a discrete corpus asset with measurable value to AI systems retrieving from it. Other Forbes-adjacent deals followed across 2024 and 2025 as the AI industry built out its licensed-source pipelines. Associated Press signed with OpenAI in July 2023. Axel Springer (Politico, Business Insider, Bild, Welt) signed with OpenAI in December 2023. News Corp (Wall Street Journal, New York Post, Times of London, Australian, Sun) signed in May 2024. Vox Media and The Atlantic signed in May 2024. Condé Nast (Vogue, The New Yorker, Wired, GQ) signed in August 2024. Reuters expanded its existing Anthropic and OpenAI relationships in 2025.

Each of these is, structurally, the descendant of what Dvorkin built. Forbes turned its content into a licensable asset class in 2010 by signaling that the publisher's distribution surface had measurable economic value. The AI licensing wave is the same logic at a larger scale — publishers monetizing the corpus the engines retrieve from. The brands that bought BrandVoice slots in 2013 were paying for retrieval on Google. The AI labs licensing Forbes, News Corp, Axel Springer, and Reuters in 2024 are paying for retrieval inside ChatGPT, Claude, Gemini, and Perplexity. Same mechanic. Larger market.

The framework

Three principles drawn from the Dvorkin receipts:

  • The content surface is the asset. Whether the surface is a search engine, a content discovery feed, or an AI retrieval engine, the value of being in the indexed corpus compounds. Dvorkin understood this in 2010. The publishers signing AI licensing deals in 2024 are operating on the same understanding fifteen years later.
  • Brands are publishers. Treat the function as a product, not a marketing channel. Forbes' BrandVoice gave brands a publishing-grade product inside a premium content domain. The brands that built proprietary content engines at the same caliber — HubSpot's blog, Salesforce's Trailhead, Stripe Press, a16z's content network — won permanent retrieval real estate. The brands that ran content as a marketing line item lost it.
  • Plan for the next licensing layer. The 2010 layer was Google. The 2026 layer is ChatGPT, Claude, Gemini, and Perplexity. The next layer is whatever retrieval system replaces them. The discipline is to keep producing content that the next-layer system will be incentivized to license. That requires named expertise, structured data, clear attribution, and consistent publishing cadence. Dvorkin's contributor network had all four. The 2026 winning publishers and brands have all four.

Lewis Dvorkin's name does not appear in most lists of figures who shaped the AI Communications era. It should. The architecture he built at Forbes between 2010 and 2017 — content surface as product, brands as publishers, expert contributors as scale, licensed retrieval as monetization — is the template the entire AI licensing market is now running on at scale. The 2013 piece called him a maverick. The 2026 receipt is that he was an early operator of the model the whole industry now uses.

AI Communications is the discipline of becoming the answer inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews. The Dvorkin BrandVoice architecture is the pre-AI-era proof that publishers and brands can capture retrieval economics by treating content surface as a product. The AI licensing wave is that same model at the next scale. The framework holds.

Ronn Torossian
Founder and Chairman, 5W AI Communications