Start for free
Back to success stories

Large-scale content management for ChatEurope

Centralized multilingual CMS and free AI chatbot powering a collaborative European news platform with large-scale content management and delivery

XWiki
Client Visit their website
Location
Europe
Size
150k users
Hosting Cloud

Industries

Journalism

Usecases

Public Platform, CMS

Services

Migration, Custom development

Requirements

Multilingual collaborative platform, AI chatbot, content management

The beginning

ChatEurope is the first chatbot dedicated to European news. Launched by 15 European partners and led by Agence France-Presse, the project brings together a broad consortium of European media organizations. It combines a multilingual and collaborative news platform, a conversational free chatbot, and coordinated social media channels to make European affairs more accessible, interactive, and reliable.

ChatEurope provides citizens with accurate, multilingual, and unbiased answers on European topics. The aim is to improve citizens’ understanding of complex issues, strengthen engagement, and build trust. It also serves as a test bed for media innovation, particularly in the use of generative AI (GenAI). All content is drawn from a curated database of verified news articles from trusted European newsrooms. The chatbot clearly cites the relevant journalistic sources in every response, enabling users to verify the information.

Press partners

Led by Agence France-Presse (AFP), the media consortium behind this project brings together prominent names in European journalism. It includes broadcasters such as Deutsche Welle (DW) and France Médias Monde, as well as news agencies like dpa and ANSA.

Publishers and specialized organizations including Agora, El País, Maldita.es, RFI Romania, and OBCT are also part of the project. Dpa and AFP subsidiaries, news aktuell and MediaConnect, handle communications.

Technological partners

At the heart of this innovative project, the chatbot was developed by the Romanian company Druid AI and operates using the language model of the French company Mistral AI. Deutsche Welle also contributes media processing through the integration of plain X for text, and soon audio and video translation.

Lastly, as the technical partner of the project, XWiki SAS, the company behind the open-source knowledge management tool, developed the news platform and the necessary integrations. The design of the website was done by Détroit agency.

ChatEurope is co-funded by the European Commission and is editorially independent, which is key to its credibility.

The context

ChatEurope emerged in response to growing challenges faced by European media and EU institutions, including disinformation, news avoidance, and low citizen engagement.

The project builds on the previous European Data News Hub (EDNH), which was technically architected by XWiki. The core components were used and further refined for the ChatEurope project.

The platform enables media organizations to create reliable datasets comprising text, infographics, interactive graphics, and video content. Thanks to a built-in interactive chatbot, citizens can ask questions about EU matters and receive clear, source-backed answers.

Needs

From a technical perspective, the project required a solution that could:

  • Provide a central platform aggregating content from multiple European media partners

  • Support collaborative editorial workflows, including shared calendars and coordinated publication planning

  • Integrate a free conversational chatbot connected to a curated and verifiable content base that delivers answers in multiple formats and languages

  • Offer a robust Content Management System (CMS) for publishing, managing, and curating content in various formats

  • Support multilingual publishing (7 languages and soon more), with integrated AI-powered real-time translations

  • Ensure responsive design, clear navigation, metadata tagging for searchability, and the potential for embedding external tools (i.e. live feeds)

  • Implement technical SEO best practices, including structured data, sitemaps, optimized URLs, mobile-friendliness, and performance optimization

  • Reuse and integrate large existing archives from the previous EU-funded project, EDNH (~5.1 million words)

The solution

XWiki was chosen because it had already developed EDNH, a proven best case practice, and its European-origin technology made it a natural fit as the project’s technical partner.

XWiki delivered a CMS web platform for desktop and mobile, which acts as the central hub for content aggregation, editorial collaboration, and chatbot-user interaction. The solution builds on the existing EDNH technical foundation. This allowed ChatEurope to be deployed faster and be cost-efficient while adapting to new functional requirements.

XWiki provided:

  1. Hosting, setup, and maintenance of the platform on XWiki Cloud (hosted in France through OVHcloud)

  2. Import and indexing of large content archives from EDNH

  3. A CMS enabling partners to upload, manage, tag, and curate content across formats (text, video, audio, graphics, data visualizations)

  4. Creation of a back office for managing the editorial aspects of home and pillar pages

  5. Native multilingual support with AI-powered translations via plain X, optimized per language pair

  6. Integration of the Druid AI chatbot, connected to curated archives and datasets

  7. Dedicated APIs exposing structured content to Druid AI for feeding and improving the conversational model

  8. Integration of the Calendar Application (Pro) to support collaborative editorial planning

  9. Daily news alerts, uploaded and managed collaboratively by project partners

  10. Responsive front end, structured navigation for 4 content pillars, metadata tagging, and search for discoverability

  11. Technical SEO setup to ensure discoverability and performance

“Really encouraging to see initiatives like #ChatEurope leading the way. This is exactly the kind of responsible AI we need, transparent, fact-checked, and built to inform.”
U
User from LinkedIn

The results

The implementation of ChatEurope created a centralized, multilingual hub for reliable European news. In turn, this improves access to verified information and boosts user engagement through interactive features.

Built on the EDNH foundation, the project benefited from a stable, scalable, cost-efficient base. This allows media organizations to focus on content quality, innovation, and responsible use of GenAI.

The partnership with XWiki made it possible to deliver the platform and its core features, as reflected in the following results:

  • Content integration at scale

    XWiki imported, indexed, and made searchable over 5 million EDNH words archives to power both the platform and the chatbot.

  • Multilingual support

    The native support for 7 European languages with integrated AI-assisted translation enables citizens to access and understand content in their own language and supports publishing at scale.

  • Flexible content management

    Partners can upload, manage, tag, and curate diverse content types in a single CMS interface.

  • Interactive information access

    The Druid AI chatbot integration, powered by Mistral AI, delivers sourced responses to user queries across multiple formats and languages.

  • Consistent adoption of ChatEurope

    On average, the platform receives 1,500 visits and 8,900 unique page views per month, with users from up to 95 countries.

“XWiki was a natural partner for ChatEurope. Their experience in the development of the previous EDNH project allowed us to avoid starting from scratch and contributed to cost efficiency when setting up our website. As a provider active in open-source development and a defender of European technological sovereignty, XWiki already had strong experience with European projects and serves as a technological partner for the DE FACTO, BROD, and ADMO hubs.”
Read the success story

Other success stories