Paris, France – January 10th, 2025 – Kairntech, specialized in natural language solutions, and Agence France-Presse (AFP), among the world’s top three international news agencies, today announced their technical partnership in the field of AI & semantic enrichment. AFP deploys Kairntech within its editorial console worldwide.
AFP covers world news, putting thousands of news items online on a daily basis in English, French, Spanish, German, Portuguese and Arabic. To increase the productivity and consistency of such large volumes of content processed, AFP uses an innovative editorial application for journalists.
The editorial application integrates the Kairntech technology and makes it available to journalists worldwide in English, French, German, Spanish, Portuguese and Arabic. The flagship Kairntech AI language solution supports journalists on the fly to approve the following elements before submitting a new article:
- The IPTC Media Topics classification (IPTC is the global standard for media metadata, containing up to seven layers of detail) leveraging Kairntech’s patent pending semantic footprint fingerprint technology that allows a single model to operate in multiple languages.
- The extraction of entities such as persons, locations and organizations leveraging custom built AI models as well as internal and Wikidata knowledge bases.
This semi-automated process improves the overall editorial system productivity by making news items immediately available with relevant contextual information and thereby improving the quality of search. In addition, editors-in-chief benefit from human-in-the-loop features to continuously improve the quality of the language models used: editor feedback is automatically taken into consideration and used to enrich in real time AFP knowledge base and retrain the language models.
“Kairntech is delighted to accompany AFP on this project, for which we will deploy the latest version of Kairntech platform and put our semantic fingerprints technology into production,” says Vincent Nibart, COO, Kairntech.
“At AFP, we are constantly striving to improve our editorial processes. Thanks to Kairntech named entities recognition and categorization technologies, we can offer our journalists, but also our customers, more powerful and relevant search and navigation tools” says Louis-Cyrille Trebuchet, Chief Technology and Information Officer, AFP. “AI and Language technologies in particular are high on our agenda for continued innovation in this field, but in a trusted and well-governed manner.”
Kairntech is an advanced AI language solution designed for extracting information, classifying and summarizing documents and question answering. Kairntech Studio is used to build complex natural language pipelines combining different components all the way from document conversion to output formatting. These pipelines are then integrated through a REST API within existing business applications, powered by a scalable Kairntech API Server.
About AFP
AFP is a global news agency delivering fast, accurate, in-depth coverage of the events shaping our world from wars and conflicts to politics, sports, entertainment and the latest breakthroughs in health, science and technology. With 1,800 journalists spread across almost every country, AFP covers the world 24 hours a day in six languages.
About Kairntech
Kairntech is a French AI and NLP startup from Grenoble, created in 2019. The team has a long joint history of developing and applying AI software to demanding industry challenges leveraging the latest language technologies including Generative AI.
For media inquiries, please contact
Kairntech
Vincent Nibart
COO, France
vincent.nibart[a]kairntech.com
Media Contact
Vincent Nibart, Kairntech, 33 0625680812, [email protected], www.kairntech.com
SOURCE Kairntech
