Artificial intelligence is rapidly finding its way into newsrooms, but one of the biggest concerns among publishers remains the same: how do you automate content creation without losing your publication’s editorial voice?
That’s the problem a new startup is attempting to solve – Wonder Journalist.
Rather than simply generating articles from prompts, the platform creates an AI model tailored to each newsroom, learning the writing style of individual journalists and editors before automating content production across articles, summaries and social media.
The idea is straightforward. A journalist conducts an interview or records notes, uploads the transcription and the system produces a draft article written in the publication’s own style. It can then generate accompanying social media posts and other derivative content, reducing the repetitive tasks that often consume editorial teams.
Individualisation
Unlike generic AI writing assistants, the platform attempts to preserve consistency by training dedicated models for each publication and, where required, for individual journalists.
“We check the voice, the style, the mechanics and everything,” founder Daniel Munioz explained, describing a system that creates separate editorial profiles for each client.
Behind the scenes, the company combines large language models with its own infrastructure, storing client-specific editorial information separately while using Microsoft Azure AI services as part of the overall architecture. According to Munioz, the language model itself is only a small part of the platform, with much of the engineering focused on managing editorial workflows and maintaining consistency.
The software is designed to remain collaborative rather than autonomous. Journalists can edit articles, request different editorial angles and ask the AI to rewrite stories without starting from scratch.
While many AI content tools target individual creators, Munioz believes the greatest return on investment lies with organisations producing content at scale. Newsrooms, agencies and media companies can potentially automate much of the production process while maintaining their established editorial identity.
Beta Trials
The company is currently running beta trials with a small number of media organisations and is seeking additional pilot partners. Rather than offering off-the-shelf software, the team proposes paid pilot programmes with agreed return-on-investment targets before moving to longer-term contracts.
Future development may extend beyond article generation.
Munioz is also exploring partnerships around AI-assisted content discovery and editorial curation, enabling publishers to identify emerging topics and organise information before journalists begin reporting. Such tools could help editorial teams spend less time monitoring breaking news and more time producing original reporting and analysis.
As publishers continue searching for practical applications of AI, the emphasis is gradually shifting from replacing journalists to reducing repetitive editorial work.
If that trend continues, the next generation of newsroom AI may be judged by how faithfully it preserves the voice of the publication using it.
