On a daily basis, we use natural language as a means of communication with other human beings, thanks to our capacity to understand, process and use words (verbal communication) and body language (non-verbal communication)…. but not the machines!
This is why the technologies developed by DAVI allow machines to understand and reproduce the form, language and emotions of humans to restore a natural communication pattern between them and their human users.


Our conversational engine benefits from the latest technological advances in natural language processing (TALN / NLP).
Our conversational engine uses machine learning technologies, boosted by deep neural networks (Deep Learning), thanks to our work on Word Embedding.

It allows the machine to understand the meaning of a text or voice statement, unstructured, misspelled and ambiguous in a field of expertise and to link it to the content of structured and specific responses.

Our technology thus provides the answers expected by your users, quickly and easily, whatever the wording of their question.


Beyond words, human beings are social beings, who feel and express emotions that they use daily to communicate. This is why DAVI has chosen to meet their social needs by integrating a socio-emotional dimension in human-machine interactions.

Our socio-emotional engine thus makes it possible to assign social and emotional skills to machines, which complement the technical and “trade” skills linked to professional expertise.

Our technology revolves around a system of reasoning on emotions combined with models of expression of social behavior. The engine reasons on the emotional state of the user through his verbal and / or non-verbal communication allowing him to bring an emotional touch by enriching the “trade” responses. Coupled with the real-time 3D engine, this Emotional AI technology detects social intentions in the verbal responses of the machine and triggers automatically and synchronously, non-verbal behaviors adapted in conversational agents.

DAVI is one of the very few players in the sector to have solutions and skills in affective computing, social psychology and ergonomics to be able to give life to sensitive and social virtual agents.

The Emotional AI gives personality to your digital assistants!


DAVI’s real-time 3D engine is a plugin-free, multidevice animation engine for the web made possible by WebGL and other rendering and animation technologies.

Originally, this engine was developed by our technical teams to allow conversational agents to reproduce and exploit, in real time, the codes of human non-verbal language (Emotional AI). It is also used to manipulate or interact with an object such as a product or a scene.

The 3D data is rendered directly on the display device, to allow loading imperceptible to the naked eye and smooth animation. The DAVI animation engine therefore allows a high degree of interactivity: the management of dynamic animation scenarios in real time without script, interactivity with the object, immersion possibilities …

DAVI’s real-time 3D animation engine offers the possibility of giving life to real 3D conversational agents or of animating an object or a scene, in three dimensions according to the user’s manipulations!

Alone or combined, these technologies can be easily integrated into a wide range of natural language applications such as virtual assistants, chatbots, conversational interfaces, voice assistants, callbots, etc.

R&D efforts, in particular around the semantic web (web 3.0) and in partnership with the main artificial intelligence laboratories, have enabled DAVI today to have solutions and skills in automatic processing of natural language and emotions for bring real digital experts to life.


A sustained effort: for years, we have engaged a quarter of our teams in applied research. Our work is supported by university partnerships. This effort allowed us to implement many applications in the field of conversational intelligence.
patent filings
of turnover invests in R&D each year


1st scientific interest group dedicated to Animated Conversational Agents

After many years of collaboration with the Paris-South University, DAVI has definitively moved to DIGITEO in the heart of the Saclay area in order to improve its relations with research laboratories.
It is in this context that DAVI is at the head of the ITHACA (Human Interaction-Animated Conversational Agent) Scientific Interest Group within the framework of a unique partnership contract with the CNRS artificial intelligence laboratory LIMSI. CNRS is the 1st public research organization in Europe and the 8th worldwide)

This Partnership contract enables DAVI to:

  • Benefit from the expertise of very high level researchers in the field of AI: Céline CLAVEL, Sophie ROSSET and Nicolas SABOURET.
  • Conduct research with PhD students under Cifre contract entirely dedicated to DAVI’s industrial subjects.
  • Generate new knowledge at the border between basic research and industrial applications.
  • Take up scientific and technological challenges on technical, emotional and relational aspects.

« Collaboration with DAVI is important for the activity of LIMSI. We want to play a central role in the problem of intelligent animated conversational agents. This is why DAVI is installed in its premises in the heart of the University to consolidate these relationships ».


Sylvie Retailleau, Paris-South University President