The Daily Briefing with AI Assistance
Each morning, Michelle Seo receives a report from Chattiana, her AI intelligence agent, Muse. This report outlines her agenda and task list, while also curating current topics from the headlines of various newsletters, complete with links to their sources. “As an introvert, I’m challenging myself to initiate more conversations,” reflects Seo, a public relations consultant based in San Diego. She explains that the newsletter keeps her informed about daily news and provides prompts to help ease awkward silences during Zoom calls.
Seo is far from alone in using AI to consume news. This trend has become so prevalent that some media outlets are now producing content aimed specifically at AI agents rather than human readers. This manifests in various forms, from AI-optimised feeds to articles structured for agents to generate personalised content.
Jacquelyn Cameron, revenue director at Axios, observes, “We believe the media ecosystem is changing. We see a decline in the open web, a reduction in searches, and a rise in hyper-personalisation, while traditional media is becoming increasingly significant in how people consume news.”
Media Outlets Adapting to AI Trends
Major publishing houses are responding to this shift in diverse ways. Many, including Condé Nast, the parent company of WIRED, have signed agreements allowing AI platforms to display their content. Media outlets are implementing generative search engine optimisation (GEO) strategies to ensure their articles appear in AI response queries. Last year, AIOSEO, an SEO plugin for WordPress sites, introduced a feature providing websites with LLM.txt files to assist AI in discovering and interpreting their content.
Conversely, some have taken a more defensive stance. The New York Times filed a lawsuit against OpenAI and Microsoft in 2023 for copyright infringement, asserting that these companies unlawfully trained models to generate news reports. The case remains ongoing.
However, what Axios and a handful of other publications are attempting signifies a more profound transformation concerning how media operates and who its stories are intended for. “We have always said we want to be where the audience is,” Cameron notes.
Introducing Axios Direct
During the Digiday Publishing Summit held in September, Axios announced the forthcoming launch of Axios Direct: feeds designed for AI agents. “We believe a significant portion of news consumption will shift to AI systems, natural language management systems, and devices,” Cameron states. These feeds are described as a more efficient means for agents to access Axios content, streamlining narratives to prioritise structured data, akin to a Bloomberg terminal rather than a series of articles.
The first two channels of Direct cater to corporate clients. The initial channel, set to launch this autumn, will allow AI agents of Axios Pro Deals subscribers to access content related to mergers and acquisitions. The second channel, aimed at training complex internal language models, is slated for release next year.
Starting with feeds for corporate clients is a logical approach, as corporate budgets tend to be more accommodating of experimental initiatives, particularly those that might yield even a slight competitive edge. However, Axios plans to develop a third feed next year aimed at consumers, providing individual subscribers with direct reports from Axios through their agents. All three feeds will resemble RSS feeds, lacking aesthetic appeal as they are not designed for human viewing but rather adopt a format of standardised text lines.
Every’s Innovative Approach to AI Integration
While Axios is condensing its content to accommodate AI, Every, a six-year-old software and media company, is pursuing a different strategy. They anticipate that their subscribers will treat AI as an extension of themselves. Dan Shipper, founder of Every, underscores that the key distinction between Every’s content and traditional articles lies in the inclusion of guiding questions whenever possible. This enables readers to test concepts independently. For instance, an article on a new model includes a series of questions that readers can copy and paste into AI, demonstrating how the model responds to those queries and evaluates the provided information.
Certain types of articles, such as guides, also feature buttons allowing readers to “Read with Claude” or “Read with ChatGPT.” Clicking on these buttons prompts the AI with a message: “Start with a concise summary: a bullet-point paragraph. Then, offer to delve deeper into the topic.” Writers include extensive details in their articles without concern for losing readers’ attention, knowing that agents will select the most relevant snippets.
Expanding Content Formats for AI Interaction
Two years ago, Every launched its first “extensible” article: an in-depth report accompanied by a database containing information such as interview recordings, news articles, and YouTube videos. Readers could consult a chatbot to seek additional information from the database. Recently, Every published an 8,000-word article about Codex, which subscribers can explore in various formats. Thanks to the database, subscribers’ agents can customise new content based on their interests, whether that be an article, podcast, or video. One reader might prefer a bullet-point summary, while another might request a brief video featuring interview snippets.
Shipper reflects, “As a writer, you have to make decisions about what to include in the story you tell. And the question always arises: are those the right decisions? Or are they the right decisions for a particular reader?” With AI, journalists are liberated from these choices.
Challenges and Opportunities in Monetising AI Content
As media organisations recognise AI agents as a new audience, another question arises: can publishers genuinely profit from this? Rasmus Kleis Nielsen, a communication professor at the University of Copenhagen, believes that while the market for literary agents is developing, it remains immature and highly uneven. Nielsen points out that significant questions persist regarding pricing and how AI agents will interact with content in the future.
Cameron and Allison Murphy, operations director at Axios, remain optimistic about leveraging AI as a tool for audience analysis. According to the executives, Axios’s concise and highly structured style has already made it a favourite among AI for answering queries. “Time is of the essence for crawlers,” Murphy asserts. “They don’t want to read three-quarters of the page to find the main idea. So, that already benefits us.”
However, does it truly benefit Axios for AI to crawl their content? Even if readers do not visit Axios’s homepage, the visibility of the brand itself may be advantageous, Nielsen argues. This mirrors the approach publishers take when creating content for social platforms like TikTok and Instagram, which rarely drive users back to their websites.
Overall, publications face the challenge of monetising specific types of content when a chatbot can provide instant answers. People are only willing to pay for information that other publications (and large language models) do not offer. At Every, this translates to a detailed guide on AI and its applications; for Axios Pro Deals, it means quick access to information that could transform the market. A full subscription to Axios Pro Deals costs $2,499 annually, while an annual subscription to Every is priced at $288.
Shipper views that $288 as a subscription for a human reader, alongside their agents. “Perhaps they don’t want to read all your material each month,” Shipper notes. “But if the subscription includes the agent, they might want their agent to know so they can say, ‘Hey, this is important.’” Murphy adds that Axios’s starting point is acceptance: people are already using AI to absorb information, and more will do so in the near future.
“I also try to diminish some of the existential dread by thinking that this isn’t the first time this has happened in media,” Murphy concludes. People have been informed through social media, short videos, and podcasts. “It’s not really about AI as an audience,” she emphasises. “It’s about AI as a tool that helps us reach our audience.”