Aqsone started in 2014 from the vision and conviction that Artificial Intelligence will be the fourth industrial revolution, after mechanization, electrification and communication.
Artificial Intelligence, which is still very underdeveloped today, will profoundly transform certain parts of our lives and revolutionize mostsectorsoftheindustry. It is up to us to use it wisely, respectingbothhumankindand the planet.
a useful AI
Our goal is to make organizations more efficient, by anticipating risks and identifying opportunities for growth.
Together, let’s enter the Digital Revolution and empower people to build tomorrow’s world!
Aqsone is constantly growing and this is only the beginning. Data Science is to companies what the digital transformation was 20 years ago.
Our ambitions and convictions on the market potential make us confident and determined.
Do you have questions or would you like to discuss Data Science?
In an increasingly competitive job market, retaining talent has become a strategic issue for companies. Being able to identify the talents that want to leave the company is essential in order to be able to retain them.
Aqsone has responded to this issue, as part of the HR Data Lab of a major industrial group, by developing predictive attrition models that identify these target employees and thus take the necessary measures to reduce the attrition rate and anticipate the risks of leaving. The avoided turnover of talent allows organizations to avoid losing the human and financial investment they have made.
Promote internal mobility
Recruiting new external employees in a tight labor market can be a long and expensive process. Moreover, professional development has become one of the major motivations for employees. So internal mobility is a serious solution that many companies are trying to promote.
In order to meet these challenges, Aqsone has developed job recommendation models using machine learning based on employees’ skills, open positions and historical mobility movements. Employees can thus be offered new functions within their company, increasing their professional fulfilment.
Predictive maintenance of mechanical or electronic systems
All systems tend to degrade over time, but companies often do not have information on different states of degradation, and are unable to prevent the various breakdowns that could occur. Most of the time, maintenance activities are carried out at regular intervals, without taking into account the state of each system, and in a reactive way, to repair a failure.
Thanks to state-of-the-art machine learning technologies, Aqsone offers a tool based on Machine Learning methods, which makes it possible to predict the number of cycles remaining before a system failure. These indicators allow companies to predict failures and perform maintenance operations only when necessary.
Anticipation of IT risks
Information systems supporting the development and manufacturing of complex industrial programs must be maintained and updated for decades. They must be capable of supporting the increasing pace of program development and thus the growing volumes of data. They are often not designed to handle such data flows. One of the major objectives of IT Departments is to anticipate risks at the interface and application level, to prevent the development and manufacturing chain from being jeopardized.
Aqsone has implemented tools based on innovative Machine Learning and advanced visualization technologies that make it possible to identify which machine in the information system is at risk of failure. The result is better visibility and projection capacity on IT risks.
Detection of orders at risk of delay
One of the key activities of the logistics function is to ensure that orders placed arrive on time, thus anticipating the risks of having orders delivered late. If it is possible to detect or predict order delays, then the order manager can take action in anticipation instead of suffering delays and chasing after suppliers.
Aqsone has developed a predictive model based on a history of various data (orders, receipt slips, forecasts, supplier records, article data, etc.) and different sources (SAP, CRM, etc.).
The predictive model estimates the risk of delay of new orders, and proposes root causes. The logistics function can now take preventive actions internally or with its suppliers.
Optimization of raw material purchasing
Buyers of large groups have access to several thousand article references from dozens of different suppliers, and must comply with multiple contractual constraints (market share, qualification, etc.). The number of possible combinations being almost unlimited, finding the best solution to minimize the total purchase price is manually impossible.
We have developed a solution based on constrained optimization calculations which proposes the best global offer in terms of supplier and price, while taking into account all contractual clauses. The gains displayed are of the order of 10%.
Analysis of workplace accidents and development of a corrective action plan
Safety at work is a key aspect on a production or assembly line. The analysis of the root causes of accidents is a difficult challenge to overcome, as the amount of data to be integrated is so large. Aqsone has developed a solution that integrates a large amount of data (HR, production, prevention and safety, location, etc.) to gain a better understanding of accidentology and identify action levers to reduce the risk of accidents as much as possible. Automatic language processing analyses were carried out in Python, in particular to identify the circumstances of the accidents, the tools involved in these accidents, and the areas of the body impacted by these injuries.
Checking the wearing of safety equipment (helmet, gloves) through videos
Workplace safety awareness plans are most effective if they are well-targeted. To this end, Aqsone has developed an algorithm capable of analysing surveillance camera footage to anonymously measure the wearing of safety equipment, such as helmets or gloves. The solution implemented consists of a neural network model that has been re-trained to identify people wearing or not wearing a helmet.
Protection of personal data
Employers are now subject to certain obligations towards their employees with regard to the protection of personal data. One of them is to inform employees about processing operations and possible violations thereof.
Aqsone has developed a tool based on advanced deep learning technologies that allows documents containing personal data, such as e-mail addresses, personal identifications, banking or health data, to be identified on the company’s various servers. These developments were carried out in python language, and use different techniques of Machine Learning, Deep Learning, and automatic language processing, as well as character recognition (OCR). The company thus has full control over the content of its directories, whether public or private, and is in compliance with the general regulations on data protection (RGPD).
Detection of expense claim fraud
The management of expense claims in the majority of companies is inefficient and fails to counter fraud, despite fairly significant investments (manual workload, dedicated tools, etc.). Beyond the financial amounts, which are very important, it is the image and ethics of the company that are at stake.
Aqsone is developing data analytics techniques to improve fraud management and detection by directly analyzing receipts using advanced Machine Learning and Deep Learning algorithms. There are numerous use cases: categorisation of scanned tickets, identification of duplicates, extraction of data from the scans, analysis of the location of the expense claim, verification of deductible VAT, etc. The technologies involved range from NLP (Natural Language Processing) to Neuron Networks.
The solutions implemented allow a significant reduction in the costs of expense management for the company and an improvement in performance.
Thank you for helping us to quickly explore and visualize the data of our A320 business flows, in order to model and predict the behavior of the information systems. Your competence and pragmatism have been valuable assets in clarifying and validating our concept.
A special thank you for the support and for the innovative and creative sharing in the definition and implementation of an industrial data flow simulator.
Global Market Forecast
Pioneering sustainable aerospace is the purpose of Airbus. By implementing modelling techniques, such as clustering or optimisation, and developing visualisation tools like R Shiny app in our project, you’ve actively contributed to this transition of the aviation industry towards a more sustainable world.