Axios and Every start publishing news for AI agents

WIRED opens with Michelle Seo, a San Diego-based public relations consultant who gets a daily briefing from Chattiana, her Muse AI agent. The briefing lists her agenda and to-do list and pulls water-cooler topics from the headlines in her newsletters, with links to the sources. Seo says she is "kind of challenging myself as an introvert to start more conversations." She is far from the only person reading news through an AI assist, and WIRED says the habit has become common enough that some outlets now produce content that treats AI agents, not human readers, as the intended audience. That can mean feeds optimized for AI or pieces structured so that agents can build personalized articles.

The article sets this against other publisher responses. Many publishers, including WIRED's parent company Condé Nast, have signed deals that let AI platforms surface their content. Outlets are working on "generative engine optimization" (GEO) so their articles show up in answers to AI prompts. Last year AIOSEO, an SEO plug-in for WordPress sites, launched a feature that gives sites LLM.txt files, which help AI discover and interpret their content. Others are defensive: The New York Times sued OpenAI and Microsoft in 2023 for copyright infringement, alleging they illegally trained models on its reporting, and that case is still ongoing. What Axios and a handful of others are trying, WIRED says, is a more foundational shift in who a story is for.

Axios announced Axios Direct, feeds designed for AI agents, at the Digiday Publishing Summit in September. Chief revenue officer Jacquelyn Cameron says: "We do believe that a significant portion of news consumption will move into AI systems, LLMs, and devices." She describes the feeds as a more efficient way for agents to reach Axios content, cutting back on narrative to prioritize structured data, closer to a Bloomberg terminal than a series of articles. The two initial feeds are for corporate clients. The first, launched this fall, gives the agents of Axios Pro Deals subscribers access to merger and acquisition content. The second will act as a source for training in-house large language models and launches next year. A third feed, which Axios plans to develop next year, will be consumer-facing and send Axios reporting straight to an individual subscriber's agent. All three will resemble RSS feeds: standardized lines of text, not meant for human eyes. WIRED notes that starting with businesses makes sense because corporate budgets tolerate experiments better than individual subscribers do.

Every takes a different route. The six-year-old tech media and software company announced a $2 million funding round last year from investors including LinkedIn founder Reid Hoffman. Founder Dan Shipper says its writers put prompts into pieces wherever possible, so readers can test ideas themselves; an article about a new model, for example, carries prompts a reader can paste into an AI to see how the model answers and evaluates input. Some articles, such as guides, have "Read With Claude" and "Read With ChatGPT" buttons that open a prompt asking for a tight one-paragraph, bulleted summary and an offer to go deeper. Writers can pack in detail, WIRED says, knowing agents will pick out the relevant parts. Two years ago Every launched its first "extendable" article, a long-form feature with a database of reporting (interview recordings, news articles, YouTube videos) that readers could query through a chatbot. Its recent 8,000-word article about Codex goes further: drawing on the database, subscribers' agents can build a new piece to fit their interests, whether an article, a podcast or a video. One reader might want a bulleted summary, another a short film of interview clips. Shipper says: "People are gonna do crazy, crazy stuff with this."

The money question is unresolved. Rasmus Kleis Nielsen, a professor of communication at the University of Copenhagen, says the market of publishing for agents is "emerging" but immature and "extremely uneven," with major questions about price points and how agents will interact with content. Cameron and Axios chief operating officer Allison Murphy are optimistic. They say Axios's short, highly structured style has already made it a favorite for AI to pull from. Murphy: "Time is money for crawlers. They don't want to have to go three quarters down the page to find the main point." Whether that crawling helps Axios's bottom line is the open point; Nielsen says brand visibility can be positive in itself, much as publishers make content for TikTok and Instagram that rarely brings viewers back to their sites.

WIRED adds that publishers struggle to monetize content when a chatbot can hand over answers, and that people pay only for information others, including LLMs, lack. At Every that is in-the-weeds AI guidance; at Axios Pro Deals it is fast access to information that could move markets. An all-access Axios Pro Deals subscription costs $2,499 a year; an annual Every subscription costs $288. Shipper sees the $288 as a subscription for a human reader plus their agents: if the subscription covers his agent, he says, the agent can tell him when something is important. Murphy says Axios starts from acceptance, since people already use AI to absorb information and more will, and that this is not the first such shift in media, after social media, short-form video and podcasts. In her words: "It isn't really AI as the audience. It's AI as a tool that helps us reach our audience."

Key facts

Axios announced Axios Direct at the Digiday Publishing Summit in September: RSS-like feeds of structured text for AI agents. The first, for Axios Pro Deals subscribers' agents and covering mergers and acquisitions, launched this fall.
Two more Axios Direct feeds are planned for next year: one to train in-house LLMs, and one consumer-facing feed that sends Axios reporting to individual subscribers' agents.
Every, whose founder is Dan Shipper, puts copy-paste prompts and "Read With Claude" or "Read With ChatGPT" buttons in articles; its 8,000-word Codex piece lets subscribers' agents reshape it into an article, podcast or video.
Prices cited: $2,499 a year for all-access Axios Pro Deals and $288 a year for Every. Rasmus Kleis Nielsen calls the market for agent-directed publishing "emerging" and "extremely uneven."
Axios executives say its short, structured style already makes it a favorite for AI to pull from; Allison Murphy frames AI as a tool for reaching the audience, not the audience itself.

Why it matters

Publishers have so far answered AI mostly with licensing deals, GEO tactics, LLM.txt files or lawsuits. Axios Direct and Every's extendable articles go a step further: they design the product for an agent that reads on a person's behalf. Axios's Jacquelyn Cameron cites a shrinking open web, shrinking search and growing hyperpersonalization as the reasons. If a meaningful share of news reading moves into AI systems, as Cameron says Axios believes, then what a story looks like and who pays for it both change.

Who it affects

Publishers deciding between defending content from AI and formatting it for AI. Subscribers who use agents to digest news, like Michelle Seo with her Muse agent Chattiana. Corporate clients, the first target of Axios Direct (Axios Pro Deals subscribers' agents now, in-house LLM training next year). Writers, since Dan Shipper says Every writers pack in detail knowing agents will select what matters, and the article's own phrasing is that journalists "no longer choose" what goes into a story.

How to use it

For readers, the practical entry points are in the article. Every's pieces include prompts to paste into an AI, and some guides have "Read With Claude" or "Read With ChatGPT" buttons that open a prompt asking for a one-paragraph bulleted summary and an offer to go deeper. Axios Pro Deals subscribers can have their agents use the M&A feed, which costs as part of a subscription priced at $2,499 a year for all-access; an annual Every subscription is $288. For publishers, the examples show two models: trimming already brief, structured content for machines (Axios), or supplying a source database and prompts that let a reader's agent reshape the story (Every).

How solid is it

This is a reported feature built on interviews with executives at the two companies, one outside academic (Nielsen) and one user (Seo). The Axios Direct details come from the company's own executives, and only the first feed has launched. The claim that Axios's style makes it a favorite for AI to pull from is attributed to Cameron and Murphy, not independently measured. The article gives no traffic, revenue or subscriber figures for Axios Direct or Every's agent-oriented content.

Risks and caveats

The business case is unproven. Nielsen says price points and the way agents will interact with content are still major open questions, and he does not say agent publishing is profitable. WIRED asks directly whether AI crawling helps Axios's bottom line; the only upside offered is brand visibility, which Nielsen likens to posting on TikTok and Instagram where viewers rarely return to the site. The legal fight is also unresolved: The New York Times' 2023 suit against OpenAI and Microsoft is still ongoing. Note too that the consumer-facing Axios feed is only planned for next year, and the second feed has not launched.

“It isn’t really AI as the audience. It’s AI as a tool that helps us reach our audience.”

— Allison Murphy, Axios chief operating officer