Marie Bering recently became Stibo DX’s Director of Product Management. With 20 years of experience spanning journalism, IT, and product development - including work in both publishing and broadcast in Denmark - she brings a unique perspective to the evolving media landscape. In the following interview with WAN-IFRA’s Dean Roper, Bering shares her insights on how AI is reshaping journalism - from newsroom workflows to editorial ethics - and explores the deeper questions facing media in the age of automation.
“It’s interesting to be at Stibo DX facing all these profound changes that we’re seeing right now – some of the biggest transformations that the industry has seen, in my opinion,” she says. “What is journalism in the age of AI? Existential questions. It’s no longer about fixing just the tooling around it – it’s considering a much larger perspective. And everybody is in need of really good support for that journey.”
In this interview with Dean Roper, WAN-IFRA’s Director of Insights, Marie Bering explains how she thinks about AI and how Stibo DX is incorporating it into the platform and tools they develop for news media worldwide.
Dean Roper: How would you describe the progress of the newsrooms you’ve worked with in adopting generative AI? Do you feel that journalists themselves have a clear vision for how AI can enhance their editorial work and storytelling?
Marie Bering: We see a lot of forward-thinking innovation out there. We see an industry that has the courage to embrace some of these opportunities and has embraced them. Of course, there are many questions about how to do this. Everybody’s learning along the way. But what we’re seeing is – and I love this because it’s not just about AI and adopting AI – that it is about how to do it in a way that supports the core business and the core values of journalism.
The ethics and the contemplation of how to do this are always at the forefront. What we’ve seen is a lot of sandboxing and a lot of piloting in the last two years. We’ve reached a point now where everybody’s looking at how to scale, how to go beyond the experiment. Everybody has rapidly realized this is not something that exists on the sidelines of publishing. It’s not just an assistant. It is a foundational new approach, and I see journalists looking for ways of preserving what makes journalism unique and valuable. When ChatGPT came out, I actually predicted a lot more caution. But I have seen a lot of newsrooms taking great leaps.
Dean Roper: Yeah, we’ve seen the same thing. The industry’s often – fairly or unfairly – criticized for not moving quickly on new tech and big changes, or especially change within. But with AI, I’ve been surprised by how they’ve really jumped on it.
Marie Bering: The change we’re seeing this year is that the first iteration of AI is already becoming a legacy experience, when everything was centered around the chatbots and generating something with AI that was useful. But now, with agents and the kind of orchestration that you can apply, AI can operate at scale, perform sequences of tasks, and then it becomes something else. We are focused on following the trends and also looking at scalability.
The adaptability and the kind of work streams that flow underneath all of this matter too. However, we are still very focused on the checks and balances that need to be in place no matter how good or how advanced the AI is. What we’re doing at Stibo DX now with CUE Autopilot is looking at a much more holistic perspective: seeing how we can make everything completely automated, but ensuring that no matter where you are as a journalist, editor, or newsroom manager who wants to have that extra check – we will make sure you have the option.
Dean Roper: Somebody at our Congress said that AI itself perhaps hasn’t started a new transformation, but it’s accelerated every aspect of business in terms of change, strategy, tooling, culture, audience, relationships and development. Do you feel like the buy-in from the top editors is behind that?
Marie Bering: We can compare it to when social media came and how the media industry handled that disruption of the distribution model. That was a big disruption. But it was still a discussion within the boundaries of how we produce the content. With the rise of AI, we’ve seen some opposition, but pretty quickly partnerships have been established. There’s a duality with this. On one hand, the media is protecting the business model of journalism and making sure they get their customers and subscribers to use their content. On the other hand, they are looking for ways to rapidly scale AI to support and make their production efficient. Those perspectives seem to coexist.
But with AI, the discussion goes deeper, back to basics. What if AI ends up developing ideas and doing the research? What’s left? What is journalism at its core? And that’s why at Stibo DX we are not just chasing the trends, but are focused on what makes journalism unique and valuable in a society. How do we support trust and transparency – the things that you cannot compromise. There will always be a need for editorial control. “We want to control this.” Or “This will never be automated.” Where publishers are seeing AI’s real value.
Dean Roper: With everything developing and changing so quickly in terms of development with LLMs or other companies, how do you strike that balance as a provider in terms of your vision that you need and how you support them, versus customer demand, where they might want changes and adopt the latest bells and whistles.
Marie Bering: That’s definitely a balance, but I think we’ve made some pretty important choices along the way. At Stibo DX, we’ve managed to approach this more platform-wide and as a system layer. So prompting in CUE Autopilot, CUE, the media enterprise platform, now is completely configurable. It is what you want it to be. So while others might have defined AI features in a more generic way, we support every media organization’s specific flavor of AI. Maybe it’s generating a headline. Maybe it’s more specific, like conforming the tone of voice to younger audiences or supporting your language in a specific way. Maybe it’s something very practical, like shortening the article.
We have built the underlying framework for managing your own prompts. We cannot define what an agent does across 100 different media companies. But we can define how to manage it, how to control it, how to make sure it works the way you intend it to. I think the scalability and orchestration that relate to managing AI is very important because the underlying technology will change, and it will only get better. We have learned that approaching this at a system level, building the infrastructure for AI, is much more important than hooking up one specific AI model. It’s aligned with the vision of the CUE platform that you can make it your own. It will suit your specific editorial needs, and it will grow with you. We do not put limitations on it. We create the flexibility. And what we’ve learned over the years is to do this by design. We make sure you can manage AI in a way that is standardized, scalable.
The question is, what kind of journalistic production we need to support in the future? We’re coming from a reality where multichannel publishing was the core problem: how do you reach as many people through as many channels with as much relevance as possible? Now we look beyond the trend of the moment and to at the shifts that are taking place in the core of content production, which is becoming much more fluid. It’s moving faster than any of us anticipated.
This report was originally published by WAN-IFRA in September 2025.
Title: Where publishers are seeing AI’s real value – so far: Strategies from 10 global case studies