For most news organizations, the adoption of artificial intelligence has followed a familiar, fragmented pattern: a few journalists using a chatbot for brainstorming, a social media manager experimenting with captions, or a data team testing a specific script. But dmg media is attempting to move beyond these isolated experiments by building a comprehensive AI foundational layer for the newsroom, designed to sit beneath the entire editorial workflow rather than acting as a standalone tool.
The initiative, centered around a suite of tools called Mail iQ, is not intended to replace the act of reporting or the nuance of writing. Instead, it targets the administrative friction—the “periphery” tasks—that often distract journalists from their primary mission of generating content. By automating the repetitive elements of filing, formatting, and distributing stories, the publisher of the Daily Mail, Metro, and Recent Scientist is attempting to industrialize efficiency across three continents.
The project began in early 2025 when dmg media established its Innovation Hub. A founding team of four initially operated in a “sandbox” environment, creating proof-of-concept products to see where AI could actually add value without compromising editorial integrity. However, moving these experiments from a controlled environment to a massive user base—including one of the world’s largest English-language news sites, MailOnline—presented significant technical hurdles.
To bridge the gap between prototype and production, the company joined the 2025 edition of the Newsroom AI Catalyst, an accelerator program run by WAN-IFRA in partnership with OpenAI. This collaboration helped the team integrate their AI tools directly into the company’s technical infrastructure, moving away from fragmented apps and toward a unified system.
The architecture of Mail iQ
At the heart of this transformation is a “multi-agentic architecture.” Rather than relying on a single, general-purpose AI, Mail iQ utilizes an orchestrator agent that manages a variety of specialized sub-agents. Each sub-agent is designed for a specific task, ensuring that the output is tailored to the precise needs of different editorial teams.
This structure allows the system to act as a proactive assistant. For example, as a journalist files a story, the system can simultaneously suggest SEO-friendly headlines, generate metadata tags, and propose the most effective URLs for the content management system (CMS). This removes the mental load of administrative filing, allowing the writer to focus on the narrative.
Crucially, dmg media has implemented a strict “human-in-the-loop” policy. The agents do not generate original journalistic reporting; instead, they analyze and optimize the work already produced by humans. Every suggestion—whether We see a style correction or a social media post—must be validated by an editor before it ever reaches the public.
Streamlining distribution and social reach
One of the most immediate impacts of the AI foundational layer has been seen in the social media teams across the UK, US, and Australia. Previously, creating a social asset for a story was a manual process that could take upwards of five minutes per post. With the social asset generation tool, that time has been reduced to less than a minute.
This efficiency has allowed the teams to scale their output to more than 300 assets per day. Beyond speed, the tool has enabled a more strategic approach to platform-specific distribution. For instance, the system now suggests specific subreddits for articles, allowing the publisher to enter communities they previously lacked the bandwidth to target effectively.
“The social teams absolutely love it,” said Chris Clemo, Director of Innovation at dmg media. “Before, it was taking them, say, 5 minutes to post something, and now it takes them less than a minute.”
Maintaining global editorial consistency
For a global media house, maintaining a unified voice is a perennial challenge. Dmg media operates across different time zones and cultures, making it difficult to ensure that every journalist adheres to the same internal style guide. Mail iQ addresses this through an AI editorial assistant that analyzes drafts and provides real-time suggestions based on the company’s specific guidelines.
By training a Large Language Model (LLM) on internal rules—such as when to spell out numbers or how to use percentage symbols—the company has created a digital guardrail for quality. Currently, about a third of the global newsroom uses the style guide assistant, with the US office acting as the primary champion for the tool before it is rolled out more widely.
To provide a clearer view of how these tools function within the newsroom, the following table summarizes the core components of the Mail iQ system:
| Feature | Primary Function | Key Benefit |
|---|---|---|
| Style Guide Assistant | Analyzes drafts against internal rules | Global brand uniformity |
| CMS Metadata Tool | Generates SEO headlines and tags | Reduced administrative filing time |
| Social Asset Tool | Creates platform-specific posts | Posting time reduced from 5m to <1m |
| Analysis Layer | Tracks internal data and live trends | Actionable, real-time insights |
| Newsroom Intelligence | CMS-integrated suggestions | Enhanced content timeliness |
The path to CMS-agnosticism
The long-term vision for the Innovation Hub is to make these tools “CMS-agnostic.” While some features currently exist as Chrome extensions or separate apps, the goal is to integrate them fully into the company’s in-house CMS. By making the tools independent of a specific system, dmg media can eventually roll out the AI layer to all its titles, including Metro, The i Paper, and New Scientist.
The scalability of this approach is a result of the multi-agentic framework. Because the foundation is already in place, adding a new specialized agent to the system now takes hours rather than days. To support this growth, the Innovation Hub has expanded its team to seven members, including the recent addition of two AI engineers.
The next phase of development will focus on fine-tuning existing agents and building new tools tailored to the specific requirements of different editorial desks. The company is currently working to fully integrate these features into the new version of its CMS, which will mark the final transition of Mail iQ from a suite of assistants to a true foundational layer of the newsroom.
We invite you to share your thoughts on the integration of AI in journalism in the comments below or share this story with your network.
Worth a look
