Founded in 2017, reciTAL is one of the first international teams to master and exploit large pre-trained language models.
As early as 2017, we used Transformers-based language models to provide document processing tools. Thanks to their pre-training on large volumes of data, these models have an innate language understanding faculty that facilitates document processing tasks. In addition to “seeing” the document, the machine can also read it and understand the meaning of words and sentences. These technologies make it possible to drastically reduce the volume of data needed to learn the task.
Whether extractive or generative, these models have brought IDP (Intelligent Document Processing) into a new era.
reciTAL now offers a complete LAD / RAD solution to exploit the power of large language models, On Premises or in SaaS.
The reciTAL team was built around two principles: excellence and diversity.
reciTAL has an international team (+10 languages spoken) of very high level (5 doctors and engineers from major schools).
Led by a cutting-edge research team (scientific committee with the University of Berkeley), reciTAL counts several major French banks and insurers among its clients, representing more than 80,000 active users.
reciTAL was a member of the G20 des Entrepreneurs 2017, won the AI Awards and the Finance Innovation favorite in 2019 and received the DeepTech label from the BPI in 2021.
reciTAL raised €3.5M with Breega in September 2020 to accelerate its development.