Le Point: testimony on the use of ContentSide Semantic Platform

Le Point improves SEO and reading time on its website thanks to artificial intelligence


With more than 80 million pages read each month, Lepoint.fr is one of the most important French news website in terms of traffic. More than 100 journalists write more than 1,200 articles every month.

In 2019, in order to modernise and improve its editorial chain, the IT Department and the Lepoint.fr Editorial Department have embarked on a digitalization project by automating, thanks to artificial intelligence, the indexing part of the project, with the aim of facilitating the work of journalists and improving the user experience.

After a study with several software publishers and the realization of a POC (Proof of Concept), Lepoint.fr chose ContentSide Semantic Platform and asked ContentSide to integrate its REST API into its editorial production system based on its document repository.

The application was integrated within a few weeks and has been operational since the end of July 2020. Today, almost all journalists are happy to use the "button" that calls the ContentSide API, thus harmonising more like this links and increasing the number of page views.

"We chose to work with ContentSide because of all the solutions we were able to test, it gave the best results, especially in terms of quality and speed, right from the POC phase. We worked as a small team, between the IT department, the publishing manager and ContentSide, to achieve a very satisfactory result. In addition, the SaaS model of the application allowed us to integrate the solution quickly and at a lower cost," says Olivier Bost, IT and industrial director of Le Point.

AI at the service of the E-journalist


Since the appearance of the first news websites, the pace of publication has continued to increase, leading to a new way of consuming information. For each article, the journalist manually integrates metadata for indexing(keywords, named entities, more like this links, a waste of time and a source of error that forces the media's editors to correct or complete the missing elements.

Today, thanks to the ContentSide Semantic Platform API, a verticalized SaaS solution that natively integrates automatic language processing features through learning(automatic classification according to the IPTC (International Press Telecommunications Council) standard, article proposal through similarity analysis, named entity extraction, etc.), the application achieves an average recognition rate of nearly 95%. The rate of relevant proposals is estimated at more than 90% for classification, 80% for the proposal of related articles and 100% for named entities thanks to the data repository provided by Le Point.

"Our API and web services interface easily and quickly with the editorial software on which journalists work, we simply added a button. With a single click, the algorithm - which integrates the documentary repository provided by Le Point - automatically fills in the fields for them and they only have to check and validate the article. This way, we eliminate these time-consuming tasks and allow journalists to focus on their job, while better valuing their work," says Arnaud Dumont, Founder and CEO of ContentSide.

The main benefits identified by Lepoint.fr are

  • 100% of published articles are read on. Articles are indexed in the system for 6 months to be linked to other publications and to offer only the most recent articles to readers.
  • Time and productivity savings for journalists and their editors, who were involved in the project from the outset, and who were able to concentrate on writing their content.
  • Reduction of the risk of errors linked to manual data management and harmonisation of "reading pursuits".
  • Optimisation of search engine optimisation through the integration of more relevant keywords.
  • Improved user experience as users are redirected to linked articles and view more articles.

"These initial results are promising and show that these AI technologies are now reliable, easily installed and usable without heavy investment, for a few hundred euros per month by a news company. After this first successful project, we are thinking about extending the technology to other applications, such as integrating paper articles or facilitating asynchronous searches of document collections: searching for publications for journalists, content feedback for the production of special issues integrating different sources," concludes Olivier Bost, IT and industrial director of Le Point.

Watch below a video workshop with Le Point explaining the business issues and the implementation of the solution:

Are you interested in this topic?