Germany’s General-Anzeiger gears AI projects to increase overall staff use
For the General-Anzeiger team in our Newsroom AI Catalyst programme, the goal was the other way around: They wanted to identify cases where AI could not only help solve something specific but also increase a wider overall uptake of AI use across their newsroom and organisation.
General-Anzeiger (GA) has a long history of being innovative. For example, the news publisher took part in our Table Stakes Europe programme not once but twice. Likewise, GA has now twice taken part in our OpenAI Newsroom AI Catalyst programme. They were also a winner of our European Digital Media Awards in 2022 for their foodie newsletter, Bonn Appetit.
With a track-record like that, it’s not a surprise that even before starting the Newsroom AI Catalyst programme, the Bonn-based publisher was already in an advanced early-adoption stage.
“We had strong AI champions, group-wide governance, and successful use cases in transcription, user-needs analysis, and AI-supported content generation,” said Marcel Wolber, Head of Digital Development at GA. “But adoption was uneven across teams and there were almost no technical bridges between our content, metadata, performance data, and AI tools. As a result, we weren’t yet unlocking the newsroom’s truly high-impact AI potential. Our hypothesis going into the programme was that a small set of well-defined, integrated use cases would scale AI adoption across the newsroom far more effectively than isolated tools,” added Johanna Heinz, who served as project lead for GA.
Interestingly, Wolber told us GA doesn’t have a dedicated AI team, but rather a task force that is made up of members who also work in various departments across the publishing house as well as their wider group (General-Anzeiger is part of Rheinische Post Mediengruppe /RPM) who come together to work on specific projects.
“The core team for this project consisted of six people spanning editorial, project management, IT, and data engineering,” Wolber said. “All members also contribute to other topics alongside their AI work, and the wider newsroom and other teams were involved through assessments and workshops throughout the project.”
General-Anzeiger’s progress was helped by RPM’s groupwide decision to roll out an AI integration platform, which has provided them with a unified foundation.
Five AI prototypes shaping newsroom workflows
In terms of specific developments, GA has built first prototypes of the following five assistants (as described by Heinz):
Newsletter Assistant: Retrieves audience-relevant articles from GA’s archive, summarises them, and drafts editorial ideas in the author’s tone and style.
Context Generator: Pulls together relevant background information from their archive to give journalists quick, well-sourced context for any topic they are working on.
Article Clinic: Diagnoses article performance and suggests optimisations based on tracking data, article data, and their user-needs model.
Fact-Checker: Verifies facts, statements, names, and spellings against their internal archive and web sources.
AI Guidelines Assistant: Makes GA’s AI policies conversational by answering compliance questions, providing example scenarios, and explaining foundational AI terms in plain language.
Since the conclusion of the AI Catalyst programme in December 2025, Wolber said the General-Anzeiger task force has continued developing these use cases, which are now in the testing phase, with “some already being trialled by colleagues in the newsroom,” he added.
Workshops, trainings help to boost staff confidence and buy-in
To prepare their staff for using the new AI tools, Wolber said General-Anzeiger ran a two-day prompting workshop for their editorial, data, and project management staff. The publishing house is also adapting their open-format monthly AI office hours into shorter training sessions that are topic-based and directly tied to specific use cases.
“The use cases themselves are not yet in daily editorial production – that’s the next step – but the workshops and early prototypes have already built confidence and accelerated buy-in,” Wolber said.
Likewise, the project team is already seeing a number of positive signs.
For example, Wolber cited the following specific early results they have seen from the project phase:
- A clear technical foundation with identified integration points and aligned tools;
- Strategic clarity through five defined high-impact use cases with shared editorial–tech ownership;
- Editorial buy-in built through workshops and prototyping;
- Data-enabled AI through the connection of performance data and their vectorised archive, which unlocks RAG-based use cases;
- Validated prototypes.
Furthermore, he noted, these use cases have been tested with General-Anzeiger’s editorial teams and other staff and have shown concrete efficiency and quality gains.
Key lessons learned: Integration and ownership
Asked about some of the main lessons they have learned so far, Wolber stressed that integration beats accumulation. “Connecting our existing data (article archive, tracking data, user-needs model) into a shared AI environment unlocked far more value than adding new standalone tools.” He also noted that editorial ownership is essential: “Involving editorial staff early through the prompting workshop and prototype testing built the trust needed for adoption.”
In addition, Wolber said, “these projects always take longer than you’d like, and they need a clear driving force that takes the lead on the issue.”
Next, General-Anzeiger will begin to roll out the use cases into their daily editorial work, Wolber said.
They are also planning to add resources that could help their editors and journalists with research, such as council information systems, city information and traffic updates, he said.
Once the use cases go into live, daily use, Wolber said they will look at “qualitative feedback from editors on usefulness and time savings, and – particularly for AI-supported online articles – quantitative performance signals such as conversions, media time, and the share of ‘ghost’ (low-engagement) articles, among others.”
The post Germany’s General-Anzeiger gears AI projects to increase overall staff use appeared first on WAN-IFRA.
