Axios Builds News Feeds Designed for AI Agents, Not People

Axios is building news feeds designed to be read by artificial intelligence agents rather than human subscribers, the company's chief revenue officer told WIRED, part of a broader bet by some publishers that a growing share of news consumption will move through AI systems instead of articles.

What is Axios Direct, and how does it work?

Axios announced the feeds, called Axios Direct, at the Digiday Publishing Summit in September, according to WIRED. Jacquelyn Cameron, Axios's chief revenue officer, described the format as closer to a Bloomberg terminal than a traditional story: structured data with the narrative stripped out, built so an AI agent can parse it rather than a person reading prose. The first Direct feed launched this fall and gives AI agents used by subscribers to Axios Pro Deals, a premium business service, access to merger-and-acquisition content, WIRED reported. A second feed, planned for next year, is meant to train corporate clients' in-house large language models. A third feed aimed at a broader audience is also planned for next year, per the report.

Why start with corporate clients instead of ordinary readers?

Cameron told WIRED that corporate budgets tolerate experimentation better than individual subscriptions do, "especially if these unproven concepts could offer even a slight business advantage." That logic explains why Axios's first two Direct feeds target business subscribers and internal AI systems rather than the general public. Cameron framed the move as a response to, not a cause of, how people already consume information. "We believe that the media ecosystem is changing," she said, pointing to a deteriorating open web, shrinking search traffic, and rising hyperpersonalization as readers lean on AI tools.

Are readers already getting news through AI agents?

At least some are, according to WIRED's reporting. The outlet profiled Michelle Seo, a San Diego public-relations consultant who starts her day with a briefing from an AI agent called Chattiana, built on the Muse platform. The agent compiles her schedule and to-do list and pulls "water-cooler topics" from newsletters she subscribes to, with links back to the original sources. Seo said the digest helps her start conversations: "I'm kind of challenging myself as an introvert to start more conversations," she told WIRED. Consumer AI agents built for similar daily-briefing tasks, including tools such as Smorgi Apps' AlterEgo, have multiplied as more people hand off routine information-gathering to software.

How are other publishers handling AI's growing role as a reader?

Axios's strategy sits alongside several other approaches, per WIRED. Many organizations, including WIRED's parent company Condé Nast, have signed licensing deals letting AI platforms surface their articles. Others are tuning content for what the industry calls generative engine optimization, or GEO, so stories surface in AI-generated answers. The WordPress plug-in AIOSEO introduced a feature last year that generates "LLM.txt" files to help AI systems discover and interpret a site's content, WIRED reported. Not every outlet has taken a cooperative posture: The New York Times sued OpenAI and Microsoft in 2023, alleging the companies trained AI models on its reporting without authorization. That litigation remains ongoing, according to WIRED.

What is still unresolved?

WIRED's reporting does not specify how Axios will price or measure engagement on feeds built for machines rather than people, or whether the general-audience Direct feed planned for next year will supplement or eventually replace Axios's human-readable journalism. The report also does not address how feeds built to minimize what Cameron calls "narrative" — the prose an AI agent is meant to skip over — will affect reporting standards or correction practices once a human editor is no longer the first checkpoint a reader encounters.

What to Watch

• Whether Axios's third Direct feed, planned for next year and aimed at a general audience, reaches consumers directly or through third-party AI agents.
• Whether other publishers build separate, non-narrative feeds specifically for AI systems rather than simply licensing existing articles.
• The outcome of The New York Times' copyright suit against OpenAI and Microsoft, still pending as of WIRED's report.
• Whether generative engine optimization practices reshape how outlets write headlines and structure stories to surface in AI-generated answers.

For readers curious how consumer-facing AI agents compile daily briefings in the meantime, tools like the mindfulAI Keyboard offer a smaller-scale look at how software summarizes and surfaces text for a single user, the kind of behavior publishers are now trying to design content around at scale.