Experts from Stibo DX, Bright Sites, Atex, and Finland’s Ilta-Sanomat, discussed some practical steps for embedding AI into newsroom Content Management Systems during the webinar, which was hosted by our WIZONE Marketplace, and moderated by Kim Svendsen, our Membership Director.
A common theme throughout the session was how AI has become the backbone of modern CMS.
Today, CMS platforms are integrating AI to support the full editorial workflow – from news gathering and production to distribution and monetisation. It’s more about serious workflow support, editorial safety, and scalability.
A key question in the session was: which AI tools should be built into the CMS, and which should stay separate?
Brian Alford, Founder and CEO of Bright Sites, said, “If an AI feature supports everyday publishing tasks such as optimising headlines, tagging, or fact-checking, it makes sense to integrate it directly into the CMS. But more strategic or data-heavy tools are often better as separate apps.”
This balance helps avoid clutter and keeps the CMS easy to use for journalists.
Kalle Pirhonen, Producer, Editorial AI at Finland’s Ilta-Sanomat added, “Not every AI function should be in the CMS because it could make things too complex for journalists. You have to be thoughtful about integration.”
Too many features clutter the CMS
Pirhonen summed up the AI journey in newsrooms as “The copy-paste evolution.”
Until two years ago, AI was mainly being used for generating summaries, translations, and SEO headlines.
“We were manually copying content from AI tools into our CMS,” Pirhonen said. “Now, AI features are more integrated, but not everything should be baked into the system. That just creates clutter.”
The focus now is on smarter integration. For Pirhonen’s team, that also means obeying strict ethical and legal guidelines.
“Any unpublished content must stay within our in-house, secure AI tools. We whitelist what’s safe. It’s not always convenient, but compliance comes first,” he said.
Language support is evolving
Sara Forni, Product Manager at Atex, introduced the idea of “fluid media” – a concept borrowed from Ezra Eeman during his speech at WAN-IFRA’s Congress in Krakow.
However, current AI-based Content Management Systems are not fluid enough to support different languages.
“Right now, most large language models perform best in English, Spanish, German, and French. African languages and some others are not well supported, but smaller, language-specific models are emerging,” Forni said.
Marie Bering, Product Manager on CMS and AI at Stibo DX, said, “We offer custom model integration, so publishers can plug in local AI models to meet regional language and legal needs.”
Options such as prompt chaining can help overcome this hurdle, but they come at a price. For now, language support is not fully integrated across platforms, but CMS developers say it can be customised based on the newsroom requirements.
Interacting with news as a conversation
Too often, systems built by tech companies don’t reflect how journalists actually work, according to Alford of Bright Sites.
“Every newsroom is unique. They have their own processes, and tech platforms should adapt to them, not the other way around,” he said.
At the same time, reader behaviour is shifting – from traditional search to interacting with news through chatbots or AI assistants. CMS developers are now exploring front-end experiences similar to ChatGPT, where readers interact with news as a conversation.
“For instance, we’re imagining workflows where editors can drop articles into a chatbot interface, and the system tailors responses for users such as topic-specific bots for elections or climate change,” Forni said.
This means CMS tools must adapt to how readers consume news.
SEO still matters, but it is changing
SEO remains important even with AI, but tactics are changing, Alford noted. It’s less about keyword stuffing and more about quality journalism.
Headlines for Google search differ from those for social or print, so tools that help manage SEO in context are valuable. Moreover, AI tools in CMS can help newsrooms optimise content for multiple platforms.
The overall goal is for AI to help journalists, not replace them.
“AI is a quality amplifier that saves time on routine tasks so journalists can focus on storytelling,” Alford said.
The best AI-powered CMS platforms are flexible and adapt to each newsroom’s unique workflow, encouraging collaboration between editorial, commercial, and technical teams.
“The challenge isn’t just using AI. It’s about using it naturally in daily newsroom routines. AI tools should be customisable, letting newsrooms keep control and maintain journalistic standards,” Bering said.