Our history

Explore the highlights of the past 15 years at Probayes

Probayes was founded in 2003 by scientists from Inria and France’s national center for scientific research, CNRS. The founders’ field of study was the modeling of complex interactions between robots and their environment. 

Probayes developed a new probabilistic programming system, the first of its kind to automate calculations performed on models and factor in uncertainty. This advance was the result of early forms of artificial intelligence based on probabilistic modeling. In addition to providing strong mathematical formalism, the approach could also handle the complexity of the world and the uncertainties of the information being manipulated. 

Back in 2003, these methods were part of a new cognitive science movement called “connectionism” that aimed to explain how the brain works using Bayesian and neural networks. Connectionism put forward an alternative to the “symbolic” movement, based on expert systems, which was beginning to show its limitations. This was the scientific environment in which the early versions of the probabilistic calculation and Bayesian modeling (ProBT) platforms were developed.

Probayes’ founders (Emmanuel Mazer, Pierre Bessière, Jean-François Miribel, Juan-Manuel Ahuactzin, Olivier Lebeltel, and Kamel Mekhnacha, currently Chief Operating Officer) held a strong belief that businesses would not only want, but need these innovations in the future. They set a course to make these new technologies reliable, robust, and scalable, so that they could be commercialized.

Probayes was founded in October 2003 as a bridge between advanced mathematics research and companies’ need to forecast and optimize in a sophisticated and complex world.



Probayes expands into machine learning.


Defense: probability

Software components combining 5D geometry and probabilitiy for the defense industry.


Operational research

Start of operational research and combinatorial optimization activities.


Marketing: data mining and predictive analysis

Probayes addresses churn with data mining and predictive analysis techniques.


Industry: optimization

New production optimization solutions for the microelectronics industry. Diversification of mathematical methods deployed on customer projects.


Defense: innovation award

MCM – ITP projects (MBDA – DGA – DSTEL) win innovation award.


Logistics & distribution: data recognition

Improved address recognition on letters and parcels.


Chatbot: automatic/natural language processing

Online postage stamp store chatbot based on automatic language processing.


Probayes founded

The first company to bring operational probabilistic tools to the market.



Probayes develops an advanced driver assistance and sensor fusion program patented in 2009.


Banking & finance: credit card fraud detection

7.5 million transactions analyzed
per day with a response time of 50 milliseconds.


DEFENSE: aircraft trajectory prediction

Aircraft trajectory prediction and missile guidance optimization earn kudos from British Ministry of Defence and French DGA.


Logistics: time-series data analysis

Innovation in time-series data analysis for gesture recognition applications. Prediction of maritime container flows.


Energy: energy efficiency for buildings

Building energy performance service based on a patented technology introduced.


A La Poste company


Probayes team grows to 50


Banking & finance

FraudIA solution released.

  • 2003 - Création de probayes

    Première société à mettre sur le marché des outils probabilistes

  • 2004 - Apprentissage automatique

    Diversification de l’activité autour de l’apprentissage automatique et Machine learning

  • 2005 - Véhicule autonome : fusion de capteurs

    Développement d’un programme d’assistance avancée à la conduite et de fusion de capteurs qui débouche sur un brevet en 2009

  • 2007 - Défense : probabilité

    Composants logiciels alliant géométrie 5D et probabilités dans le domaine de la Défense

  • 2008 - Banque : Détection des fraudes à la carte bancaire

    Analyse de 7,5 millions de transactions de paiement
    par jour avec un temps de réponse de 50 millisecondes

  • 2008 - Recherche opérationnelle

    Démarrage de l’activité en recherche opérationnelle et optimisation combinatoire

  • 2009 - Défense : Prédiction de trajectoires aériennes

    Prédiction de trajectoires aériennes et optimisation du guidage de missiles. Travaux primés par la DGA et son équivalent britannique le MOD

  • 2010 - Marketing : Data mining et analyse prédictive

    Ouverture aux problématiques de l’attrition dans le marketing avec les techniques du data
    mining et de l’analyse prédictive

  • 2011 - Logistique : Analyse de données temporelles

    Innovation dans le domaine de l’analyse de données temporelles avec des applications de
    reconnaissance gestuelle. Prédiction de flux de conteneurs maritimes.

  • 2012 - Industrie : optimisation

    Optimisation de la production pour l’industrie micro électronique. Diversification des méthodes mathématiques appliquées dans ses projets.

  • 2013 - Énergie : Réduction énergétique des bâtiments

    Développement d’une offre de performance énergétique des bâtiments (technologie brevetée)

  • 2014 - Prix meilleure innovation

    Projets MCM – ITP (MBDA – DGA – DSTEL)

  • 2016 - Filiale du Groupe La Poste

  • 2017 - Distribution : reconnaissance de données

    Amélioration de la reconnaissance des adresses sur les plis postaux

  • 2018 - Probayes intègre son 50ème salarié

  • 2018 - Chatbot : TAL / NLP

    Traitement Automatique du Langage avec lancement du Chatbot « Mon timbre en ligne » #NPL

  • 2019 - Banque

    Lancement de l'offre FraudIA