2023-05-03 06:30:45
To remain competitive, manufacturers must allow their engineers to focus on tasks with high added value such as breakthrough innovations. It is to help French companies meet this challenge that the start-up Dessia has developed software that acts as a design companion, powered by generative engineering (AI). This technology thus enables Dessia’s customers (mainly large manufacturers in the automotive, aeronautical and rail sectors) to improve the profitability of projects by reducing design times.
Incubated at Télécom ParisTech, this start-up created in 2017 benefited from a fundraising of 5.5 million euros in 2021. Explanations with Pierre-Emmanuel Dumouchel, founder of Dessia
Engineering Techniques: How can “bots” help engineers in their decision-making?
Pierre-Emmanuel Dumouchel: I’m going to take as an example the case of Renault with which we worked to define the cooling architecture. When you place your heat engine, gearbox and other elements such as air conditioning, engineers have the choice between millions of ways to route all the cooling systems, to place the heat exchangers… In this specific case, we have developed a bot which will exhaustively generate all the possible paths and which in the end propose a hundred of them which are very interesting, including the first three which are considered to be the most optimized. This is called decision support.
Was the integration of these “virtual companions” easily accepted?
Initially, our goal was to automate everything in the search for an optimal solution. However, we realized that allowing the engineer to choose which solution to adopt was crucial to ensure the acceptability of the approach. In general, we find that engineers are willing to adopt new ways of working by delegating certain exhaustive build tasks to a bot. However, it is essential for them to regain control at the end of the process.
By reducing the duration of a design iteration, your solution might encourage companies to reduce their engineering teams?
It’s one way of looking at things, but not all of the major companies with which we work in the automobile (Renault), aeronautics (Airbus) and railways follow this logic. They want to free engineers from design offices and others to be able to mobilize them on tasks with higher added value such as research and longer-term breakthrough innovation such as hydrogen propulsion. For French manufacturers, the main challenge is not to fall too far behind Tesla and the Chinese brands that are entering the market.
We must rethink the way of designing products if we want to be competitive
Yes. For the moment, we are deploying bots to our customers on isolated businesses at first. But quickly, manufacturers realize that this integration has advantages. With a reformulation of the knowledge of their different teams, the company is led to deeper reflections in order to review its way of working in the different trades and to modify its interfaces.
What new professions might appear thanks to AI?
In industry, there are already distinctions between “methods” and “operational” departments. Our customers want to evolve by considerably increasing the size of the “methods” departments and reducing the other operational part. This would be more and more with automated processes. The “methods” departments would include data scientists who would be, in a way, the “craftsmen” in charge of equipping all the automated processes. If we look at the way of working at Tesla, we see that the departments that think regarding the integration of AI are very developed.
What are your projects ?
Initially, our objective is to develop on a European scale. Our technical challenge is to improve our approaches to orient ourselves towards Low-code, so that generalist engineers, who are not experts in data science and who work in ETIs where there is little or no at all, from data scientist profiles, can use our tools to develop their own bot. By offering low-code solutions, we aim to improve the accessibility and use of our bots in companies.
1683335257
#PierreEmmanuel #Dumouchel #Decision #support #bots #engineers #focus #innovation
