As news organisations seek faster content production, CMS vendors are increasingly embedding AI within editors’ existing tools, prioritising practicality and editorial control to transform workflows without costly system overhauls.
Newsroom CMS vendors are increasingly treating artificial intelligence not as a bolt-on experiment, but as part of the publishing infrastructure itself, as editors look for faster ways to produce, package and distribute content without breaking established workflows. In a recent WAN-IFRA discussion, suppliers argued that AI adds the most value when it sits inside the tools journalists already use, rather than forcing them to move between separate applications.
The central argument was one of practicality. Tom Pijsel of WoodWing said the point of embedding AI in the CMS is to cut out repetitive switching, allowing journalists to stay inside a single environment while carrying out tasks such as headline generation, copy editing and page construction. Massimo Barsotti of Eidosmedia said standalone AI tends to create friction, while integrated tools can shorten text, turn material into tables and even generate charts from within the editorial interface. Sara Forni of Atex pointed to automated transcription and "voice-to-story" workflows that can convert raw audio and video into draft copy for review.
That shift is also widening the use cases for AI beyond summaries and social posts. Eidosmedia said publishers are beginning to look at automated pagination for print, where layouts are assembled according to editorial rules with less manual page-building. Other vendors are pitching similar ideas: Quintype promotes AI-assisted curation, image handling and omnichannel publishing; CoreMedia describes an in-context assistant built into its CMS; and Avid has tied AI to smarter search, recommendations and orchestration in a cloud-native content platform.
A second theme is that many publishers do not want to replace their core systems just to adopt AI. Forni said newsrooms cannot afford long migrations that disrupt live operations, which is pushing vendors towards layered approaches that add capability to existing platforms. Atex’s editorial layer connects with systems such as WordPress and Drupal, while WoodWing and Eidosmedia both emphasise API-first designs that allow AI tools to plug into current production setups rather than force a wholesale rebuild.
Even as automation expands, all of the suppliers stressed that editorial judgement must remain with people. Barsotti said agent-based systems may be able to link content, create sections and run tasks in parallel, but journalists still need control and safeguards. The common position across the market is that AI should be editable, reversible and subject to review, not allowed to make final decisions on story selection, tone or brand identity. That approach has become a defining principle for vendors trying to persuade publishers that AI can improve efficiency without undermining trust.
According to WAN-IFRA, the webinar drew 310 registrants from 90 countries, with most participants in senior or mid-level publishing roles. The size of the audience suggests the issue has moved well beyond curiosity. For many news organisations, the question is no longer whether AI belongs in the newsroom, but how deeply it should be wired into the systems that already run it.
Source Reference Map
Inspired by headline at: [1]
Sources by paragraph: - Paragraph 1: [2], [6] - Paragraph 2: [1], [3], [6] - Paragraph 3: [2], [3], [4], [7] - Paragraph 4: [1], [3], [4], [5], [7] - Paragraph 5: [1], [3], [6], [7] - Paragraph 6: [1]
Source: Noah Wire Services