This shift has been dubbed “liquid content” — a catch-all term for stories that can be reshaped and tailored to readers, and redistributed across multiple formats with minimal effort. 

In many ways, it’s the next step in the personalization of content.

“Traditionally, publishers create content as a finished object. An article, a video, a story. Liquid content shifts that thinking toward content as structured knowledge that can flow into different formats, surfaces and interfaces. For publishers, that’s a fundamental change,” said Marcel Semmler, CPO at Bauer Media.

What is liquid content?

The term gained legitimacy when it was included in a recent report on journalism, media and tech predictions from the Reuters Institute for the Study of Journalism published on Jan. 12.

Here’s how the report defines liquid content: “content or stories that are not static but adapt in real time based on the viewer’s context, location, time, or interaction. AI facilitates this by tailoring content to individual preferences. Requires traditional media companies to move away from authoring ‘articles’ towards more flexible atomic objects.”

Agentic AI browsers — like Perplexity’s Comet, OpenAI’s Atlas, The Browser Company’s Dia — can do this already, by reading, summarizing and acting on information inside the browser itself, pulling together content into personalized news briefings, for example.

“I think of liquid content as content that is no longer designed as a fixed output, but as something that can be reassembled and experienced differently depending on context,” Semmler said. “It moves the conversation away from pages and placements and toward systems, structure, and intent.”

OK, but what does that look like in practice?

It might all sound a bit theoretical now, but there are already some live use cases.

Perhaps the most basic example is how information in a published article appears in an interaction with ChatGPT. The information from a single story can fuel answers in ChatGPT, and generate personalized experiences for publishers’ readers.

There’s some debate over whether using AI to transform a text article into an audio briefing — something that a myriad of publishers, including The New York Times, The Washington Post, Time, and others already allow — is considered liquid content. Dmitry Shishkin, a media consultant who works with companies like Swiss-based Ringier, doesn’t think so. Instead, he considers this “multimodality.”

True liquid content means the “content can adapt to the individual user, including their preferences, user needs, context, time of day, and behavior,” according to Shishkin.

A better example of liquid content is a recent project by The Washington Post: the publisher used AI to build a pick-your-own-format news podcast, letting listeners choose the topics, hosts and length to create custom versions. Users can choose to listen to an AI-generated podcast, which selects and stitches together roughly four top stories, based on the user’s reading and listening history and topics they follow in the app, all brought together by two AI hosts summarizing the news stories and conversing back and forth. The podcast continues to change throughout the day, based on the news cycle. (Unfortunately, that initiative was plagued with inaccuracies.)

“Liquid content is difficult to make while retaining factual accuracy,” noted Clare Spencer, a reporter at generative-ai-newsroom.com. Any project using generative AI runs the risk of inaccuracies, due to the fact that the technology is designed to make predictions based on the data it’s trained on, she said. One simple way of overcoming this is having a human check the facts before publishing, she added.

Companies like Forbes, Newsweek, Time and The Washington Post are thinking about how to transform their sites to make them more personalized, and anticipate reader needs, and interact with visitors in real time, based on data like geolocation, referral sources and past reading behavior. The website experience could be flexible enough to reflect what the user wants in that moment (such as a list of bullet points on the news tailored to the user’s preferences, a video or an AI chatbot experience) — another example of liquid content.

Norway-based publisher VG recently built a news product called VGX, which is a feed of news content built from information pulled across the company and then reformatted by AI agents.

Why does liquid content matter right now?

Core to the idea behind liquid content is that articles as we know them now will be an old format one day.

“The article becomes irrelevant,” said David Caswell, founder of StoryFlow, a consultancy focused on AI workflows in news production. “It’s the end of documents… Publishers will go from sellers of documents to sellers of information.”

As readers increasingly get information from LLMs and AI answer engines, publishers need to think beyond their sites and traditional search results, Semmler argued.

“If publishers don’t adapt to that reality, their content doesn’t disappear, but it becomes invisible. It still creates value, but somewhere else. What publishers should be doing now is not chasing the next format, but investing in foundations. Structuring content more deliberately. Being clearer about intent. Separating knowledge from presentation. Building systems that allow content to be reused without losing editorial quality,” Semmler said. 

Is it difficult to do?

Yes and no.

The technical side of creating liquid content isn’t difficult, but the quality of the output can be tough to get right. As can reorganizing a newsroom to not be as focused on fixed outputs and channels, as it requires working across editorial, product and technology teams, according to Semmler.

Publishers can start experimenting with liquid content by taking something like a long-form podcast, putting it through AI tools to reformat it (such as into a video for TikTok), and checking it for quality (looking at accuracy, bias and copyright), Spencer suggested. “If you think this is good enough to share, then share and see what your audience thinks,” she said.

But do readers want liquid content? What’s the value for them?

This remains to be seen, but is ultimately the most important question behind all of this. True liquid content is in its infancy among publishers’ tests right now. But why go through all the trouble of building these experiences if it’s not what readers want?

“The collective opinion of consumers is the only thing that matters. We will find out,” said Caswell. “But a big chunk of the early evidence is the consumer response to chatbots generally — such as ChatGPT and AI Overviews. I think it’s quite clear that consumers want what they want, when they want it, in the way they want it.”

Semmler argued that liquid content can help reduce friction, speed up access to information, and make experiences feel more relevant to readers (audio for in the car, bullet point overviews in the morning, a video in the afternoon). Bauer is looking at what this means for their organization now, he added. 

“We’re not treating liquid content as a format experiment or a distribution trick. We’re looking at it as a product and platform question,” he said. Liquid content can play a strong role when readers are looking for answers, but not as much of a role when they are looking to be entertained, Semmler added.

What other considerations are important for publishers?

Before investing in making content liquid, publishers need to understand user intent, according to Semmler. Publishers also need to ensure they still retain ownership over their content, he said.

“As content flows through external systems, publishers need to be very clear about attribution, control, monetization, and brand presence. Otherwise, they risk doing a lot of work to power experiences they don’t own,” he warned.

Shiskin had a different perspective, suggesting that liquid content was a distraction from what publishers should be putting resources into now to battle the threat of AI to their businesses: focus on original, high-quality journalism that readers can’t find elsewhere and are willing to pay for.