While most news organisations are using AI tools to streamline workflows within the newsroom, an increasing number are also developing AI-powered products and features to engage their audiences in new ways.

The experiences of the two French publishers Les Echos and Le Parisien, which were shared at our Paris AI Forum, point to some promising applications in this area.

Les Echos: Focus on retention with audio playlist & evening briefings

With a newsroom of 220 journalists and up to 30 million monthly visitors, Les Echos is a major French business news outlet founded in 1908.

Despite its long history, the outlet has gone through an extensive digital transformation, with 60 percent of its revenues now coming from digital sources, according to Violaine Degas, Chief Digital Officer at Les Echos.

Aligning with a subscription-driven business model centred on individual and B2B subscriptions, the publisher is developing new, audience-facing AI solutions aimed at driving daily user habits. The goal is to “engage the subscribers, which is very key for retention,” Degas said.

One example is the text-to-speech feature “Playlist de l’actu” (“News Playlist”) that was launched exclusively for premium subscribers in October 2024. Updated twice daily, it provides audio versions of the latest top articles, allowing readers an easy way to keep up with the latest essential news on the go.

Although Degas acknowledged that the feature has seen fairly modest use so far – partly because it is only available to premium subscribers, “so the base is not huge” – many of those who do use it do so regularly.

“It’s really a driver for engagement. We are thinking of improving it, even if the volume of people using it is not huge,” she said.

Another feature aimed at premium subscribers is “18-20,” a special section of the app where the newsroom chooses the 12–15 most important articles that are highlighted between 6 pm and 8 pm every day. Launched in January, the feature has been a “big success among the subscribers, but also among new readers who are coming to the application to discover this selection,” Degas said.

One feedback about the feature has been that some clients would like to offer an English version of the selection to their foreign employees based In France.

Les Echos is now working on offering this, using generative AI for the translations. “It’s the first time we are using AI to translate something so visible, so we want the translation to be very good,” Degas said.

The team has evaluated several options for translations by testing different LLMs and using the publisher’s old articles, based on technical criteria and translation quality. Testing is still underway, but Degas said the team is closing in on the final selection. The aim is to release the feature at the beginning of 2026.

Le Parisien: auto-inserted videos and explanatory visuals

For Le Parisien, the development of audience-facing AI features has been guided by the needs of their readers.

As a generalist news outlet and with a newsroom of more than 350 journalists, the publisher reaches a large segment of the French audience. It has 20 million readers every month and a large following across different social networks.

One key area where Le Parisien has been active and is video, as the message from their readership has consistently been that “they want more video contents,” said Sophie Cassam Chenaï, Chief Digital Officer at Le Parisien.

Read more: TikTok: Le Parisien’s journey from zero to 400,000 followers

To better showcase its video content, Le Parisien has developed a system for automatic video insertion: an AI-powered tool compares the articles with the publisher’s video catalogue and inserts relevant video content in the articles.

“Today, we have one out of three articles with a video,” which is significantly boosting the viewing rates of these videos, said Cassam Chenaï.

Le Parisien is also enhancing its shopping guide vertical by using AI-generated videos to showcase featured products when creating an authentic video is not possible.

Another AI use case Le Parisien has identified is explaining stories more effectively and enhancing how its journalism is presented, particularly when there is no real video content of an event.

“It’s AI-augmented storytelling to explain to our readers exactly what happened,” Cassam Chenaï said.

She gave an example of an incident in which a cyclist was hit by a car in Paris: Le Parisien used AI to create a video reconstruction of the event. The publisher uses a “cartoony” style for such videos, to make it clear that they are generated content and not authentic recordings of the real event.

Coming up: answer engines

Both publishers also said they are working on answer engines which will allow readers to interact with their content in new ways.

For Les Echos, this will take the form of an Economic and Business Intelligence engine aimed specifically at enterprise and professional audiences. The goal is to help users save time when preparing for a meeting or gathering information on a specific topic.

“It’s a retrieval that we are going to build around our content and all the archive, so it’s more than 35 years of articles,” said Degas.

Currently in a testing phase, Les Echos hopes the service will stand out due to its reliability, thanks to limited hallucinations, a system that is fine-tuned to its own content, and energy-optimised hosting solutions. The publisher aims to launch the service during the first half of 2026.

As for Le Parisien, the outlet is building a response engine that will significantly improve its search function and allow audiences to explore the publisher’s archives in new ways.

“Readers can ask questions and have a summary as the answer, with all the sources and direct links to articles,” said Cassam Chenaï. Le Parisien also hopes to release its response engine early next year.

Answer engines are a natural step in an evolution that Degas called “moving from an ‘article-centric’ to a ‘service-centric’ approach to informing readers,” through greater personalisation in news products, for instance.

“We are convinced that we need to move beyond the article,” she said.

“We have moved to new formats like video, podcast and so on, but now it’s another move that we need to make, to more service-centric features. And that’s really the opportunity with GenAI.”