When industry talks to his machine

Non-intrusive and natural, voice guides operators and allows them to deliver information without delay to the computer system.

The voice assistant SIMSOFT Industry analyze and process information dictated by the technicians of the future RER E tunneling between Paris Saint-Lazare and Courbevoie.

The voice finds its place in the industry. Since the emergence of natural language understanding technologies, voice assistants are attracting more and more industries, who see a way to improve productivity and the exchange of information. “All industrial interested in voice, but that does not mean that there is a maturity in what they expect, tempers André Joly, the Director General of Industry SIMSOFT, French designer of industrial voice assistants. Enthusiasm, however, is real.” Hard not to see the benefits of this technology for industry. It allows the operator to be guided in its tasks and trained field, to obtain information and make up.

The underground extension of the RER E between Courbevoie and Paris Saint-Lazare is more than six kilometers long and ten meters in diameter. The tunnel boring machine used to dig such a tube is huge. The technicians who operate the must record every action performed. Once entered manually on a computer installed in the cabin, qualification task was entrusted to a voice recognition software. “The installation of this system was decided in particular to improve data analysis, Braud remembers Nicolas, the director of the Tunnel Lab BouyguesPublic works. previously, they were difficult to use. We had to spend much time cleaning before analysis” While an operator could describe a problem in five different ways, not learn a task for lack of time, the voice assistant immediately records standardized and usable data. ” This allows for a much more accurate view of the use of machinery, identify improvements and to trace the cause of problems“, said Nicolas Braud. After eighteen months of development with SIMSOFT Industry, Technology integrate all tunnel boring machines used by Bouygues and will be deployed in seven to eight?machinery worldwide.

Speech technology has not always had good press. “Few years ago, we saw morons operators with a voice command telling them what package recover in a warehouse, reminds André Joly. The challenge, not to repeat it, is to keep the technician's responsibility.” So to provide a wizard that increases the operator's abilities rather than give orders. this difference, decisive to facilitate its entry into the industry, comes from the maturity gained by the technology in the natural language understanding. The fact that the public voice assistants, as Google Home and Alexa speaker ofAmazon, have penetrated the homes strengthens sympathy operators to technology.

It blows a favorable technological and societal wind, welcomes André Joly. Technology is advancing and operators are already familiar with the voice assistants because they have one in their living room.” But the general public difference, the industry sector poses problems. “The adoption of speech technologies by industry stumbles over three constraints, says CEO of SIMSOFT Industry. The noise, which hinders understanding the instructions by the robot, vocabulary and operators of business constraints and the need to provide off-grid solutions, for reasons of security and practicality.” But the benefits of voice will make her an indispensable part of the future of the industry, believes there. “This is a very intrusive system, unlike augmented reality, that requires wearing bulky glasses. And the voice is the easiest means of communication.

Helping the operator and capture data

The supplier develops intelligent assistants tailored to various sectors. Airbus, EDF, Michelin and Thales benefit from its services. “Our primary market is the automobile, to improve productivity and reduce non-quality, says André Joly. Then come aeronautics, including for aid to understanding the documentation, and the energy sector, with the safety instructions of recalls and field experience feedback. Ensuite, we provide all the industries that use a lot of data and measures.” This is the case in the factory Ventana Arudy (Pyrénées-Atlantiques). This subcontractor of Aeronautics is equipped with a voice assistant that accompanies operators. “The technician communicates with the computer system while it performs its tasks. He may request or send information without manipulate tablet“, explains Christophe Richard, the director of business innovation. Still being improved, to suit the needs of operators and site constraints, Industry SIMSOFT the assistant should be deployed in the rest of the plant over the next six months. “In addition to helping operators, voice assistant will allow us to capture the information we do not have today, feels responsible. With these data retrieved hot, we can improve our processes and achieve a return on investment.

"Each plant has different use cases"

3 QUESTIONS A : Thibault Celier, Director of Innovation Viseo, consulting business and technology solutions provider

Where is the adoption of voice assistants industry?

They arrive gradually in the industry, and for good reason: improve productivity, automate tasks, relieve human and facilitate access to information and its transmission to information systems. In the factory, there are situations where one needs to have a free hand. There is case of actual use, perhaps more than in the general public.

What are the obstacles to the adoption of this technology?

Fear of the servo of the human by the robot is a. Few years ago, we saw people in warehouses whose acts were dictated by a voice assistant. We must ensure that the link between the robot and the human is not a relationship of domination. This worked very well for the general public because it was worn by the power of Gafa. And because they are setting up processing bricks of interoperable and open natural language. Manufacturers are more reluctant. They want their tools, outside the cloud.

What prevents Gafa sell their voice assistants industrial?

The industry has vertical processing of very specific natural language, while the general public assistants are effective in the treatment of a horizontal and general language. They are much less efficient on a specialized vocabulary, the health or safety eg. every industry, each plant, presents case of different uses. It is not the prerogative of Gafa, not specialized.

Press