Demystifying AI for News Publishers: Liquid Content

This guide is part of an FT Strategies series exploring the AI technologies reshaping news publishing. Each edition explains an emerging technology often misunderstood by journalists and editors — from AI engines to liquid content and from agentic systems to MCP servers — before outlining how it can be applied across editorial, commercial and product functions. Get in touch if you have specific technologies or terminology that you’d like us to cover.

What is liquid content?

At FT Strategies, we define liquid content as content that primarily consists of datafied components, or “atomic objects”, including quotes, events, or dates, as opposed to information that is already synthesised into discrete products, like an article, podcast or video segment (Reuters 6).

Imagine news content as clay; right now, publishers create content that is moulded, shaped, and dried into a finalised product. Liquid content offers a way of keeping this clay “wet” so it can be moulded into multiple finalised products as needed. This means that a user could specify their preferred story format depending on topic or mood; for instance, if a reader has a fifteen-minute commute to consume news, liquid content allows them to easily pull up an audio briefing that fits their available time. Eventually, experience engines and decision engines will be in charge of this moulding, autonomously shaping a news product to be personalised to an individual user’s needs.

What does it offer and why should publishers care?

The concept of liquid content becomes an attractive proposition to publishers because it allows for the low-effort generation of multiple end products that are designed for distinct audience segments, particularly when coupled with multimodal generative AI capacities. When fully realised, liquid content will fundamentally change the economics of journalism by enabling customised, high-value user experiences, which can both increase engagement and unlock new revenue opportunities.

It will also enable a story to serve multiple user needs, rather than forcing users into a single format, allowing for complete user personalisation. This should make editorial operations more efficient and impactful because, in theory, more stories will find their target audience and fewer will go unread or unwatched because it was produced in the “wrong format”.

The emergence of liquid content will also introduce new challenges. For instance, the means of measuring the efficacy or effectiveness of a story will change; if one story can be consumed in multiple formats, how will publishers know if it succeeded? There is additionally the question of keeping a human-in-the-loop for published content; if generative AI is being used to produce end content, how can publishers trust the faithfulness of interpretations?

Despite these challenges, in a landscape where audiences are increasingly consuming news through AI platforms, this flexibility to tailor content consumption is essential for news publishers to remain competitive (Reuter’s Generative AI and News Report 2025).

What is an example of liquid content — and what is it not?

In some ways, the news publishing industry already has a working model of liquid content: visual or data journalism. This kind of natively visual content, which can be displayed through interactive maps or graphs, relies on inherently structured and modular data that can easily be resynthesised in real time according to audience preferences or user-driven exploration.

Content repackaging with text-based content, however, is less established, likely because it begins with a fixed article. Because articles are not modular, they cannot be easily reformatted in the same way as data underlying visuals. As a result, these text-based outputs are often more generic, aligning poorly with user needs.

With liquid content, we apply the same logic of data journalism and visualisations to text, allowing for more user-oriented solutions.

It is also important to clarify what liquid content is NOT:

• A means of reformatting an already-complete piece of reporting, like an article, into another form of finished reporting, like a podcast. This is format repackaging, not liquidity.
• An interactive interface, like summary generation or chatbot feature, that is layered on top of an article or archive. This is a way to place AI capabilities on top of pre-existing pieces of content, rather than a natural fluidity of the original reporting.

• A way to statically personalise news websites for users, like with a “Recommended For You” section. This is a personalised distribution system, not a means of actually altering content.

Why isn’t liquid content widespread?

There are several reasons why liquid content has not been adopted en masse by news publishers.

Firstly, there are still risks with publishing user-facing AI instances without proper editorial oversight. Though many publishers are experimenting with refiguring their content into different formats — the Financial Times’ AskFT chatbot is one example — many digital media experts warn against the viability of fully-enforced liquid content systems. True personalisation via liquid content means it would be impossible for humans to review every form of content reproduction. “The technical side of creating liquid content isn’t difficult,” Sara Guaglione writes, “but the quality of the output can be tough to get right” (WTF is Liquid Content?). As AI systems improve, however, content quality and versioning faithfulness will improve with them.

Secondly, liquid content requires a complete reinvention of newsroom norms, and a deliberate shift to prioritising readers’ tastes over organisational processes and habits. A recent study published by FT Strategies and WAN-IFRA found that this is already a challenge: just 32% of newsrooms have achieved alignment between their strategy and daily editorial coverage. In a liquid content-driven world, the priority must be delivering what users want to read. This non-technical component is harder to solve for, and requires newsrooms to fully embrace a user-first mentality.

How do you monetise it?

While liquid content is a nascent idea for many news publishers at the moment, FT Strategies envisages three potential revenue drivers once it is fully operational within an organisation:

1. Retention tool: liquid content will increase user engagement by formatting content in the way that the audience desires it. It will also provide scope to test non-traditional, user-interactive formats, including games, quizzes, and Q&As, that are harder to produce. Similarly, when user format preferences change, liquid content should allow organisations to adapt easily to new tastes.
2. Dynamic advertising: in creating a way for users to interact with personalised formats, publishers will develop a better understanding of how different kinds of content appeals to end users. The same logic will apply to advertising material, which will also likely undergo a liquid content evolution.
3. Content licensing: liquid content itself — datafied, but not formalised into discrete items — may also become its own kind of marketable product. Once liquid content is structured and standardised, publishing companies could open themselves up to broader data licensing deals with AI companies or pay-per-query systems with AI agents.

What does that mean in terms of publishing?

Liquid content obviously has the potential to reinvent the whole publishing process, from content commissioning to article surfacing to consumption of the content by a user. While much is still unknown, here is what it could look like:

Content Commissioning

In a liquid content-laden world, editors will not ask journalists to compose finalised end products as frequently. Rather than assigning reporters a “600 word write-through” or a “three minute explainer video,” editors will instead expect a datafied information core of a broader story that can be easily fashioned into different story formats. While that information core might still eventually be used by a human writer to create a final piece of writing, the first stop for the reporter will be to create the datafied content, not the finished product.

Similar to the transition from print to digital, this change will require a new kind of thinking in the newsroom. Journalists will reorient their priorities towards unique topic coverage and the most effective way to tell stories — rather than being constrained by existing formats — and most importantly, will always put the reader first.

SEO

Search Engine Optimisation — meaning optimising news content to surface on search engines — is frequently discussed in parallel with AI, especially given the reduction in search-and-click revenue methods for publishers as a result of AI platforms (Google SEO definition).

Liquid content can offer a means of better orienting products for SEO by boosting engagement. In particular, the lessons publishers learned from dynamic content — or showing end users specific versions of pages — apply to liquid content (Dynamic Content and SEO). The underlying concept is the same: individually reformatting content for end users to ultimately increase engagement and improve SEO surfacing.

More specifically, because of the personalised nature of liquid content, its implementation might mean that:

1. Users spend more time on a publisher’s site or return more frequently because they are guaranteed relevant content.

2. Users are more likely to interact with content or adverts on a page because it will be personalised to them and their preferred format preference.

3. Users are less likely to quickly leave the page.

4. The content on the page is more relevant or fresh because it is drawing from the latest datafied components available to it (Dynamic Content and SEO).

While this system of content signals is constantly changing, these improved metrics (like engagement, bounce rate, pages per session, etc.) have historically led to search engines ranking the site higher in their results (Dynamic Content and SEO).

GEO

Generative Engine Optimisation refers to creating content that will surface on AI platforms like ChatGPT or Google AI Overviews (GEO: Generative Engine Optimization, What is GEO for SEO?).

In publishing, GEO tends to mean more structured articles, including direct answers to questions or bullet point formats (What is GEO for SEO?). While there is still some doubt surrounding the measurable returns of GEO, studies show that organising content in these ways improves visibility on AI platforms (GEO: Generative Engine Optimization).

Fully-implemented liquid content would allow for the easy restructuring of content towards GEO principles. Due to the fluidity of content formats, news organisations could display content in this format to benefit web crawlers or agents, while still giving users the option for an alternative, more interactive style — something that publishers like The Economist have recently explained. As GEO standards evolve, liquid content will provide the foundation for publishers to easily modify how their content is displayed, consistently supporting easy integration into AI systems.

More importantly, the core principle behind liquid content is hyper-structured data behind the stories, regardless of how the front end is configured for end users. Maintaining this kind of structure will only advantage publishers for both SEO and GEO practices.

So where should you start?

Much like in the transition from print to digital news, the movement towards liquid content will require more of a shift in newsroom norms than an actual technology lift.

Editorial teams will have to reconsider their approach to content commissioning and begin creating datafied information cores rather than discrete objects for direct consumption. As highlighted earlier, this means reinventing traditional publishing processes that have placed emphasis on the published or broadcasted product. We expect that this cultural shift in the newsroom will need to be overseen by senior leadership, who will need to provide a clear mandate to editing and production teams to redefine workflows for the creation of “datafied objects”.

As Marcel Semmler, Global Chief Product Officer at Bauer Media Group, writes: “Liquid publishing requires leaders to rethink how editorial, product and commercial teams collaborate… Technology can support this transformation, but leadership must enable it” (Liquid Content and the Reinvention of Media). This kind of organisational transformation change may take multiple years for a medium- to large publisher.

Liquid content, like many of the technologies discussed in this series, is still emerging and evolving. We’ll update this article as best practices and case studies emerge. In the meantime, please get in touch if you’d like to speak with our team of AI consultants.

At FT Strategies, we assist publishers in transforming emerging AI concepts into actionable strategies. This includes developing workflow and context engines that enhance today’s newsrooms, as well as pioneering systems such as agent orchestration and experience and trust engines. Our aim is to simplify complex processes and unlock genuine editorial, product, and commercial value.

If you’re considering how to integrate AI engines into your organisation, please get in touch for support in transitioning from experimentation to implementation. We can help you build the capabilities needed to ensure high-quality, trusted journalism.