The morning routine for Michelle Seo, a public relations consultant based in San Diego, now includes a briefing from Chattiana, her AI-powered Muse agent. This personalized summary encompasses Seo’s daily agenda, to-do list, and even curated "water-cooler" topics drawn from her newsletters, complete with source links. "I’m kind of challenging myself as an introvert to start more conversations," Seo shared with WIRED, explaining that Chattiana’s briefing provides her with a sense of current events and conversational fodder for virtual meetings.
Seo’s experience is emblematic of a rapidly evolving media landscape where artificial intelligence is not just a tool for consumption but is emerging as a direct audience. This paradigm shift is so pronounced that some news organizations are beginning to tailor their content, not for human readers, but for AI agents. This adaptation manifests in various forms, from optimizing content feeds for AI consumption to structuring articles in ways that facilitate personalized summaries and deeper dives by these intelligent systems.
The Shifting Media Ecosystem: Why Publishers Are Adapting
The notion that news organizations are increasingly producing content for AI is rooted in a fundamental reevaluation of the media ecosystem. Jacquelyn Cameron, Chief Revenue Officer at Axios, articulated this sentiment, stating, "We believe that the media ecosystem is changing." This observation is supported by several concurrent trends: the reported "deterioration" of the open web, a decline in traditional search engine traffic, and a surge in hyper-personalization, with Large Language Models (LLMs) playing an increasingly significant role in how individuals access information.
This confluence of factors has prompted major publishers to explore new strategies. Many, including WIRED’s parent company, Condé Nast, have entered into agreements allowing AI platforms to access and surface their content. Simultaneously, the industry is actively developing "generative engine optimization" (GEO) strategies, a conceptual successor to traditional Search Engine Optimization (SEO), aimed at ensuring articles are prominently featured in AI-generated responses to user prompts. Tools like AIOSEO, a popular WordPress SEO plugin, have introduced features such as LLM.txt files, designed to assist AI in discovering and interpreting website content, further signaling this shift.
However, this adaptation is not without its friction. The New York Times, in a landmark move, filed a copyright infringement lawsuit against OpenAI and Microsoft in 2023, alleging that these companies had illegally trained their AI models on the newspaper’s copyrighted reporting. This legal battle, still ongoing, highlights the complex copyright and intellectual property challenges arising from AI’s insatiable appetite for data.
Axios Direct: Serving the AI Agent
Amidst these developments, Axios is pioneering a more proactive approach, treating AI agents as a distinct audience segment. The company announced at the Digiday Publishing Summit in September the forthcoming launch of Axios Direct, a suite of feeds specifically designed for AI agents. "We do believe that a significant portion of news consumption will move into AI systems, LLMs, and devices," Cameron explained, emphasizing the need to be "where the audience is."
Axios Direct aims to provide a more efficient conduit for AI agents to access Axios content. Cameron described these feeds as prioritizing structured data over narrative, drawing a parallel to the functionality of a Bloomberg terminal rather than a traditional collection of articles. This approach acknowledges that AI agents process information differently than human readers, requiring a more streamlined and data-centric format.
The initial rollout of Axios Direct targets corporate clients. The first feed, launched in the fall, grants AI agents of Axios Pro Deals subscribers access to merger and acquisition content. A second feed, slated for launch next year, will serve as a training dataset for in-house large language models. Looking ahead, Axios plans a consumer-facing feed for individual subscribers in the coming year, enabling their personal AI agents to receive curated Axios reporting. These Direct feeds are conceptualized as visually unappealing, text-based formats akin to RSS feeds, optimized for machine readability rather than human aesthetic appreciation.
Every: Empowering the AI-Augmented Reader
While Axios is optimizing its content for AI consumption, the media and software company Every is adopting a different philosophy. Founded six years ago and having secured $2 million in funding from investors including LinkedIn founder Reid Hoffman, Every positions AI as an extension of its human subscribers. Dan Shipper, Every’s founder, explained that their content often differs from conventional articles by incorporating prompts directly within pieces. This allows readers, and by extension their AI agents, to actively test concepts and explore information interactively.
For instance, an article discussing a new AI model might include a series of copy-pasteable prompts that readers can feed into an AI to observe its responses and evaluate its input handling. Furthermore, certain article formats, particularly guides, feature interactive buttons such as "Read With Claude" or "Read With ChatGPT." Clicking these buttons triggers pre-written prompts instructing the AI to provide a concise summary followed by an offer to elaborate, effectively personalizing the reading experience. This strategy allows Every’s writers to include a wealth of detail without the traditional concern of overwhelming human readers, confident that AI agents can efficiently extract the most pertinent information.
Every’s commitment to this AI-integrated approach extends to its "extendable" articles, a concept introduced two years ago. These long-form features are accompanied by a comprehensive database of source material, including interview recordings, news articles, and videos. Readers can query a chatbot interface to access supplementary details from this repository. A recent 8,000-word article on Codex exemplifies this, offering subscribers the ability to have their AI agents generate new content formats – articles, podcasts, or videos – tailored to specific interests, from bullet-pointed summaries to compilations of interview clips.
Shipper articulated the philosophical shift: "As a writer, you have to make choices about what goes into the story you’re telling. And there’s always a question about, are those the right choices? Or, are those the right choices for a particular reader? With AI, journalists no longer choose." This sentiment underscores a potential liberation for content creators, allowing them to focus on in-depth reporting and analysis, while AI handles the customization and delivery.
Monetization in the Age of AI: A Developing Frontier
The fundamental question for publishers remains: can they generate revenue from this emerging AI audience? Rasmus Kleis Nielsen, a professor of communication at the University of Copenhagen, described the market for publishing for AI agents as "emerging" but "extremely uneven." Significant uncertainties persist regarding pricing models and the long-term interaction between AI agents and published content.
Despite these uncertainties, executives at Axios express optimism. Cameron and Allison Murphy, Axios’s Chief Operating Officer, believe that Axios’s inherently short-form, highly structured content format naturally lends itself to AI consumption. "Time is money for crawlers," Murphy stated, explaining that AI agents are designed for efficiency and prefer content that presents key information upfront, a characteristic that already benefits Axios.
The brand visibility gained from AI crawling its content, even if it doesn’t drive direct traffic to the Axios homepage, can still be beneficial. Nielsen draws a parallel to publishers creating content for social media platforms like TikTok and Instagram, where the primary goal is brand exposure rather than immediate website visits.
The challenge of monetizing content that can be summarized and synthesized by chatbots presents a significant hurdle for publications. The value proposition for readers, and now potentially for AI agents acting on their behalf, increasingly lies in information that is unique and not readily available through LLMs. For Every, this translates to in-depth guidance on AI itself and its practical applications. For Axios Pro Deals, it’s the exclusive and timely access to market-moving information, justifying its annual subscription fee of $2,499. Every’s annual subscription, priced at $288, is framed as a subscription for both the human reader and their AI agents, recognizing that the value extends to the agent’s ability to curate and flag important information.
Murphy views the current situation not as an existential threat but as a natural evolution of media consumption. She points to the historical shifts in how people access news – through social media, short-form video, and podcasts. "It isn’t really AI as the audience," Murphy concluded, "It’s AI as a tool that helps us reach our audience." This perspective frames AI not as a replacement for human readers but as a powerful new channel through which publishers can connect with and serve their audience, both human and artificial. The journey of adapting to this new reality is just beginning, promising a future where content creation and consumption are inextricably linked with the capabilities of artificial intelligence.
