Login
Login is restricted to DCN Publisher Members. If you are a DCN Member and don't have an account, register here.

Digital Content Next

Menu

InContext / An inside look at the business of digital content

Inside TIME’s rollout of its TIMEAI interactive agent

TIME isn’t just licensing its journalism to AI platforms; it has built its own agent to keep users on-site, deepen engagement, and prove that trusted archives can power differentiated AI experiences.

April 16, 2026 | By Esther Kezia Thorpe – Media Reporter@mediavoicespodConnect on
-TIME's agentic AI interface to show media use of AI-

TIME has been an industry leader in pursuing content licensing deals with companies like OpenAI and Perplexity. These ensure that the publisher’s content is cited and attributed, as well as securing revenue. 

But TIME is not just relying on AI companies citing their work. They have also developed and rolled out their own tool, TIMEAI. This is an interactive agent that has been trained on its 103 year archive in order to enrich the user experience on-site.

Time to train

TIMEAI, which appears on the site as a toolbar overlay, was initially tested at the end of 2024 with the annual reveal of TIME’s Person of the Year, Donald Trump. The toolbar was also put live on the three prior Person of the Year honorees; Taylor Swift, Volodymyr Zelensky and Elon Musk.

“There were several hundred articles about all four of them that we trained [the AI] on,” explained Mark Howard, Chief Operating Officer at TIME. “Person of the Year is our single biggest editorial event of the year. So it’s not like we put it on some back catalogue just to see if we could get some basic data. We put it on our most prominent release.”

The tech used was a small language model rather than a large one (LLM), so it was more straightforward to deploy. But it still needed training on TIME’s editorial styles, and guardrails added. TIME’s Editor in Chief Sam Jacobs and some of the newsroom team were brought in to help train TIMEAI in terms of style and tone.

“We had to, given the personalities involved, make some decisions about how it was going to handle questions that had nothing to do with anything we had written about,” Howard said. Even though there is a lot of publicly available information on these public figures, they decided they didn’t want the AI to use content outside of TIME. “It’s not a complete picture when somebody wants to converse, but we made a lot of those trade-offs in a very contained way,” he added.

The version deployed in December 2024 enabled users to translate the article, summarize it in various lengths, chat with it, and even speak to it using voice and audio in 13 different languages. Rather than just having one type of function, Howard said that being able to mix and match requests – for example, to ask for an audio summary then go into detail via text – sets this AI tool apart from many others on the market.

A wider rollout of TIME AI agent

The initial launch was a success. Howard said that the original toolbar was kept live while the team worked on bringing more TIME content into the AI. This wasn’t a straightforward task.

“Because of the way media companies operated over the decades, we had [content] in five different databases going back to the 1920’s, some of them were just PDFs of magazines that needed indexing and digitizing,” he explained. “Now it’s trained on well over three-quarters of a million pieces of content.”

Since November 2025, the TIMEAI agent has been visible on almost every piece of TIME content. Newly published pieces are also indexed in near real-time, ensuring that the answers given are up-to-date.

The team have also been experimenting with ways to get users to interact. The agent has a number of preset prompts to show how sophisticated queries can be. These can range from, ‘Try reading this article out loud’ to ‘Debate AI’s impact on humanity’ using TIME’s reporting and opinion archives.

The results so far have been encouraging. Users who engage with TIMEAI are 139% more likely to return to TIME when compared to those who don’t. Those using the AI agent also double their time spent on the TIME website.

Now, the focus is on encouraging those who may not be as familiar with AI tools to engage. Howard said that this may include putting more instructional overlays in, or feeding more preset prompts into the toolbar. But they are cautious about the extent to which people are routed into the full AI experience, rather than the standard web page. 

For now, Howard sees TIMEAI as an opportunity to increase engagement, rather than a revenue generator. “If we can get people to engage, every surface represents a monetization opportunity,” he added. “But the focus has been on the product itself in the beginning.”

A trusted agentic AI source

Given TIME’s licensing deals with other AI engines, there is a question about why users would come to TIME, rather than ask an AI which pulls in a wider range of sources, including TIME itself.

Howard was emphatic that trust is something the other AI tools don’t yet have, but TIME does. “A lot of people grew up with TIME, knowing and trusting that the red border stood for something, and there was a level of quality that is behind it,” he explained. “We want to carry that forward.”

ChatGPT and other answer engines do cite TIME content, but alongside a wide range of other sources, including sites like Reddit. These answers aren’t always trustworthy, and Howard said that there is a high burden on the consumer to deconstruct what is accurate and what isn’t.

“We think [TIMEAI] can be a trusted source of information to try and help explain the world in a way that TIME can, because of this vast archive,” he said, pointing out that users will only encounter TIMEAI if they’re already on the TIME site. This already signals a level of intent and interest in TIME’s reporting.

“It’s a way for us to pull our archives, our history, our heritage and our brand equity forward in a way that we can present more through an AI-based interface and experience.”

Howard is also clear on the role AI has to play in TIME journalism going forward. “Our journalists get information that only humans could by speaking to their sources and doing the work,” he emphasised. “AI can never do anything like that. It’s not intended to do anything like that. 

“Its purpose is to take the great journalism and then be able to make it accessible to those who want audio briefings, roundups and other formats to be able to better understand everything that we’re publishing.”

Walking the tightrope between using AI tools to help reporting, and letting them take over is something a number of publishers have fallen down on in the past months, with potential long-term implications for audience trust. 

Interactive agents present an opportunity for quality publishers to surface decades of reporting to add context and enrich the user experience. TIME’s approach is cautious, and they are wary of detracting from a carefully-designed web experience. But as similar tools roll out rapidly, readers will become more familiar with what they can do and how to interact with them. TIMEAI’s multi-feature functionality will set it apart as a leader.

Liked this article?

Subscribe to the InContext newsletter to get insights like this delivered to your inbox every week.