Did you miss INMA’s fabulous Media Innovation Week in Europe last month? We had an array of speakers on AI, talking not only about what their experiments are but also why they are undertaking them — and what challenges they face.

One theme that repeatedly popped up was personalisation. “Personalisation is now expected by the user,” as Rober Zilz, head of data at Germany’s Kölner Stadt-Anzeiger (KStA), told conference attendees.

For customisation, Switzerland’s Ringier AG is playing with GenAI in many ways. The company has four chatbots running in parallel on different technology stacks. It has also created text-to-voice versions of articles with the cloned voices of journalists and is working on improving its storytelling by generating scripts, better suited for audio than text articles being read out, said Chief Product and Technology Officer Bernd Volf. 

Ringier decided any product that it builds would focus on three aspects: relevance for users, high quality, and uniqueness. 

“Uniqueness gets more and more important — we don’t want to copycat or just show data like Google does. We want to make something unique out of it and also increase the relevance there,” Volf said.

Where does one start? “We have a lot of data, and we need to produce a lot of content … especially when we talk about hyper-personalisation,” he said. 

Ringier ended up building a localised weather assistant that was launched in Slovakia a few weeks ago. Why weather? “It is very interesting for our users and you have a tremendous good amount of data for a long period of time,” Volf said.

Using GenAI, the weather assistant produces about 90 articles per day for 82 precise locations. Within the first two weeks, 7% of daily users were reading it consistently, Volf said. Ringier expects it to generate an increase of about 5% in pageviews.  

Next, Ringier plans to roll it out across 19 markets, with deeper, one-to-one personalisation, and consider monetisation. 

Another popular GenAI trend, which you have read about here before: chat products. 

This one comes from Italy’s RCS Mediagroup, publisher of newspapers such as Corriere della Sera.

“Our starting point was a challenging one: How can the product support the growth of 650,000 subscribers?” said Daniela Buoli, head of digital product. 

The news brand decided to offer the AI assistant in an app that is exclusively for Corriere della Sera subscribers. “The app ecosystem was the best way to develop a direct relationship with our users,” Buoli said. “We tried to develop an assistant that truly matters to our readers … . The audience no longer needs to go and browse the Web to find valuable answers.”

Readers received personalised answers, article summaries, and reading suggestions on different verticals. And if it cannot answer their questions, it connects them with a human expert.

Another objective was to create new touch points between the content and the readers.

Corriere followed two principles while building the product. The first was that it was focused on subscribers to promote meaningful loyalty and long-term engagement. The second was that it must be relevant and reliable — grounded in Corriere’s trusted archive, with relevant answers sourced from over 30,000 articles. 

The brand wanted to build “a powerful tool to match editorial priorities with the real interests and real questions of our audience,” such as which topics they care about and what they do not understand.

One hurdle: audience education. “People don’t know how and why to use an assistant to inform themselves,” she said.

Volf had a similar finding on GenAI: “The hardest part is really not building it. It’s getting people to use it.”

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .