Bonnier News transformed its print production through automation

By Paula Felps

INMA

USA

For years, Benjamin Peetre worked on “trying to make nice printed pages for the readers to enjoy.” Today, as senior print developer at Bonnier News Local, he’s still doing that — but now he has some help from robots.

During a recent Webinar presented by the INMA Generative AI Initiative, Peetre explained how Bonnier News has transformed its print production process and leverages AI-driven tools to produce more than 40 newspapers across Sweden and Finland.

The initial idea for automating print production wasn’t just about technology; it was about rethinking how these titles could be produced together in a streamlined, centralised way, Peetre said.

But first, they needed the right tools.

“When you talk to people at an executive level about print automation, they get a special look in their eyes and they envision an empty office with just one desk and one big button, and you only have to push the button once a day and all the papers will be produced,” Peetre said. “But it’s not that easy. You still need to have a lot of professionals to control automation.”

The fully automated page layouts are indistinguishable from layouts done by humans.
The fully automated page layouts are indistinguishable from layouts done by humans.

Trials and test runs

Bonnier first attempted to go live with print automation in 2022, but soon discovered the process wasn’t yet good enough: “It was an unreliable outcome of the print automation. Some pages were really good and fit to print right away, but some days you got no pages that you could print without extensive manual editing of them.”

Rather than abandon the idea, Bonnier News pressed pause, gathered what it had learned, and made two key changes: It redesigned all newspaper templates to be automation-friendly, and then waited for an improved version of Naviga’s AI. When the updated technology was ready, it was tested using the new templates.

“The result was so good that we thought, OK, we can move on now and automate everything,” Peetre said. “The big question at that point was, if we have a really good robot that can make pages really fast, how do we use it in the best way to get the most out of the automation?”

The company put a lot of thought into that question and ultimately came up with a new process, he said.

Working together, but separately

It was essential to keep the print production process separate from editorial production, Peetre emphasised: “The reporters are the journalists. They’re writing articles only for online publications. They don’t write articles with a specific character count to fit in a specific space on a specific page in print. They just write for publications.”

The newsroom CMS, called Abbe, handles article creation and online publishing. Once the articles are created, they’re exported to Naviga’s print automation platform, driven by an AI brain named Flow. (Internally, it’s affectionately referred to as “the robot.”)

The platform is driven by an AI brain named Flow.
The platform is driven by an AI brain named Flow.

While the robot chooses layouts and image placements and builds full pages, it doesn’t touch the article content, headlines, or captions. It simply pulls together the best versions of articles and visuals to create polished, printable pages.  

Next, print planners — trained editorial staff — select which articles will appear in the next day’s print edition. They review and clean up content (like trimming long headlines or removing duplicate photos) and feed it into the automation system. Once ready, they hit a button and the robot generates a full print layout in five to 10 minutes.

The output is editable in Adobe InDesign if needed, but many pages are ready to print without changes. The process separates digital publishing from print entirely — reporters focus on digital, while print planners put their attention on high-quality, automated newspaper production.

Print planners select which articles will appear in the next ’s print edition and feed it into the automation system.
Print planners select which articles will appear in the next ’s print edition and feed it into the automation system.

Not just about the tools

Adopting automation at Bonnier wasn’t just a technical project; it also required a significant organisational shift. Rather than retrofitting automation into existing workflows, Bonnier News restructured its production model to support automation from the ground up.

These changes, Peetre said, were crucial. “If you just have your old organisation and put a new technology in the middle of it, you won’t get that much out of the automation process. If you make adaptations to the organisation without automation, you can get a lot of efficient improvements. But if you do both the organisation and the print automation, you get much more out of it.”

Bonnier News relaunched its automation initiative in March 2023 with 24 smaller newspapers; by spring of 2024, it was onboarding its Finland newspapers. The plan is now to bring the last of Bonnier News Local’s largest newspapers into the system.

“It’s been quite a fast journey from when we started to when we got it to work as good as we wanted it to,” Peetre said.

Next, Bonnier News is exploring the automation of ad placement, which is currently still handled on the old system. He said one of the greatest validations of the use of automation lies in the fact that readers didn’t notice the change. Even editors acknowledged the automated pages were indiscernible from manually built pages.

“In the end, it’s all about what the readers think about us and our papers,” Peetre said, adding the company didn’t announce that it was using a robot to build pages and didn’t receive any feedback from readers when the change occurred.

“That was a good thing for us,” he said. “It was a relief that it was working as seamlessly as it did.”

About Paula Felps

By continuing to browse or by clicking “ACCEPT,” you agree to the storing of cookies on your device to enhance your site experience. To learn more about how we use cookies, please see our privacy policy.
x

I ACCEPT