KEY POINTS
Nordwest-Zeitung (NWZ) is a regional news daily headquartered in Oldenburg, Lower Saxony. In addition to its online and mobile platforms, the paper produces 10 local print editions, each carrying up to a dozen dedicated local pages alongside shared national and international content. Its weekend title, Nordwest Sonntagsblatt, is published across seven local editions.
Despite heavy investment in digital channels, print remained a significant source of advertising and subscription revenue for NWZ — making the efficiency and quality of its print production operation a business-critical priority.
Across a typical 24-hour cycle, NWZ publishes approximately 120–150 unique print pages. Until recently, producing those pages depended on 18 print channel managers who spent several hours each afternoon manually constructing the layout for each local edition — one of the single largest components of the paper’s staff costs.
NWZ had operated the Eidosmedia edition-management platform since 2010. The system already provided template-based pagination and automatic story reformatting for digital and print. Yet the task of populating pages with editorial content still required substantial manual effort from channel managers. Multiplied across 10 editions and a daily publication cycle, this represented a formidable workload.
For its online editions, NWZ had already deployed high-productivity curation tools capable of handling a large and diverse content portfolio efficiently. Extending a comparable level of automation to the print workflow had been a strategic objective of management for several years.
Working with Eidosmedia, NWZ implemented an AI-powered page automation system built around a pagination engine developed by Canadian technology company Sophi.io — and fully integrated into the existing edition-management environment.
The system is currently in full production and is used to lay out around 1800 pages a month.
The stories to be included in the automated pages are prepared in the usual way. In addition to the tags applied by the author to guide online publication (priority, section etc.) the print channel manager also assigns special design features for the story. This serves the pagination engine as a guide to headline size and other parameters.
At the same time, the print channel manager selects which pages of the edition will be filled with the candidate stories. Any stories already present in the pages are ‘pinned’ to prevent them being disturbed by the auto-layout process.
The pagination engine is accessed as a cloud deployment. Before sending the page content to the engine, the Eidosmedia edition management system calculates the needed space for each variant of the article, including subheadings and captions, as well as the dimensions of a selection of headline variants.
These parameters, together with the page text and media content, are then sent to the pagination engine in Jason format.
The pagination engine uses the story parameters and the layout rules it has extracted from its training data to create an optimized layout for each page of the section. These are then returned to the edition management environment in Jason format.
After a short interval (under two minutes for a complete edition) the content is laid out in the selected pages and displayed in the page plan.
A critical design principle is that the AI-generated layout is fully editable. Channel managers can apply any standard design adjustment — copyfitting, image resizing, spacing corrections — using familiar tools. This distinguishes the solution from competing platforms that return finished PDF pages staff cannot modify.
Layouts can be updated at any time as new or revised stories become available. Pages that are complete and approved can be ‘frozen’ to exclude them from any subsequent repagination pass.
Before deployment, the Sophi.io engine was trained on a corpus of machine-readable model pages, story structure descriptions, and a classification of editorial content components.
A supervised training phase followed, in which the engine generated candidate layouts, received corrections from editorial staff, and progressively extracted the design rules underlying a publication-quality page.
The results have been substantial and measurable:
Freed from the time pressure of manual pagination, channel managers can now focus on what matters most: the editorial and visual quality of the finished page. The boundary between story editing and page editing has blurred productively — staff now have the capacity to actively update and improve content at the page stage, not just place it.
Later print deadlines — made possible by the speed of automated layout — mean that breaking news can be reflected in the print edition more reliably than before.
The efficiency gains extend to special and one-off print products. New formats, themed supplements, and reader-specific editions previously carried significant design overhead. With AI handling the bulk of standard pagination, staff can redirect creative effort to front pages, flagship supplements, and new print formats — work that genuinely benefits from human judgment and editorial instinct.
The NWZ implementation has established a validated model for AI-driven print pagination in European newspaper publishing. The same Eidosmedia–Sophi.io approach has since been adopted by multiple news-media groups across Germany and France.
Page Automation has been adopted by multiple news-media groups in Germany and France.
Read the press release.
An AI-driven print layout system implemented by Nordwest Zeitung with edition-management provider Eidosmedia, automating the pagination of print edition pages.
Based in Oldenburg, Lower Saxony, Germany, NWZ publishes 10 local print editions with up to a dozen dedicated local pages each, plus seven editions of its weekend title Nordwest Sonntagsblatt.
Print layout was highly labour-intensive: 18 channel managers spent several hours each afternoon producing around 120–150 unique pages per day — a major share of total staff costs.
NWZ had used Eidosmedia since 2010 for template-based layouts and automatic story reformatting. However, filling pages with content still required substantial manual work by channel managers.
Channel managers tag stories with print design features and select pages for auto-fill. Eidosmedia calculates space requirements and sends content to the Sophi.io cloud engine in JSON format, which returns an optimised layout — typically within two minutes.
Yes. Channel managers retain full access to standard design tools for copyfitting, picture resizing, and other adjustments — unlike some competing systems that return static PDF pages staff cannot edit.
The engine was trained on machine-readable model pages, story structure analysis, and story component classifications. It then went through a supervised phase — generating layouts, receiving corrections, and extracting the rules for suitable page design.
Print layout staffing has been cut by nearly half and pagination costs reduced by around 65%. The system now handles approximately 1,800 pages a month.
Yes. Channel managers now spend more time on editorial and graphical quality control. Later print deadlines also allow more up-to-date content in breaking-news situations, and freed-up time benefits front pages and special supplements.
Yes. Page automation using the same approach has been adopted by multiple news-media groups in Germany and France.
Eidosmedia is a global supplier of advanced content-management and digital publishing systems.
Its products are used by large news-media groups for print and digital publishing.
Customers include business dailies The Financial Times and The Wall Street Journal , as well as generalist news publications like The Times of London, The Boston Globe and Le Figaro .