2024-07-31 05:30:00
In January, Schneider offered all of its teams a “Promptathon in the box” toolbox, so that each entity could develop sessions adapted to its business lines. A promptathon is planned in North America, another for the HR team, and a third is in the works. “They allow us to identify use cases, to innovate, to be more productive, continues Jean-Côme Renaudin. Value is often found by users.”
In a similar vein, Thales and Stellantis are organizing shorter “prompt parties.” After learning how to address a generative AI, a group of employees questions it on a topic. “You give people the tool for half an hour, they test it, and two weeks later they call us to tell us what they want to do with it, Guillaume Calfati, generative AI evangelist at Stellantis, told a conference. Today we have 1,000 projects without IT people [technologies de l’information, ndlr] behind. Everyone creates their own ChatGPT and becomes a maker!”
Vinci Construction completely locked up its executive committee for three days to acculturate its members. “Many are engineers, they understood that the tool presented a real strategic interest, says Cédric Lefrançois, the digital HR director. We also wanted to show them that there were already 40 use cases in the group.” The executive committee, convinced, launched working groups in all subsidiaries to find out how this new technology could transform HR, finances, construction sites, etc.
An emergency for HR services
The arrival of ChatGPT in November 2022 had the effect of a bomb: “a real revolution”, “a stunning effect”, “we can’t keep up”, HR departments report. When they looked into the subject, companies understood that their employees were already using ChatGPT without telling them. According to an Ifop-Learnthings survey, more than half of employees who use AI at work do so in secret from their employer. With the risk of letting confidential information slip or misusing the tool by not checking their answers. It was necessary to take matters into their own hands, define a global strategy, and quickly set up training. Starting with acculturation for everyone. Start-ups rushed into this emerging market. “For the past few months, there has been a big boom in generative AI. It’s a huge wave,” confirms Béatrice Gherara, the co-founder of Kokoroe, which has been offering training in new digital learning technologies for ten years. While these were previously aimed at a few hundred employees of a company, they now reach tens of thousands. “This is the first time that the subject has not been taken up by training managers, but by business leaders, explains Beatrice Gherara. Generative AI is transforming business models, leaders are seeing the value.”
For its part, the start-up Mendo counts among its clients EY, Vinci, GRDF, Hasbro, Groupe Rocher. “They start by acculturating everyone, then train each profession – HR, marketing, development, finance, etc. – in the use cases specific to it, before harmonizing and sharing practices,” says Quentin Amaudry, the general manager of Mendo. Within the Rocher Group, owner of the Yves Rocher and Petit Bateau brands, 650 employees have been using the Mendo solution on their computers since January, at their own pace. “Generative AI makes it possible to create value across all professions, not just IT, so we have chosen to bring everyone on board,” testifies Siegfried de Preville, the group’s leader in safety, health and well-being at work. He has created professional communities that share uses and libraries of prompts. Training managers now use AI to format program sheets, HR departments to write job offers. Logistics and maintenance have also got involved. Marketing and communications departments will follow specific training courses to learn how to generate images.
At Schneider Electric, acculturation doesn’t just happen through promptathons. Specific webinars for each profession have been organized and tens of thousands of people have already followed Kokoroe’s training. Everyone can now use AI to write an email, summarize a meeting, translate a text. A more specific target, 12,000 top managers and leaders benefit from enhanced training. “We need to take them from understanding the tool to being able to produce change,” Jean-Côme Renaudin emphasizes. The third audience is made up of “Use case contributors, product sponsors. They need to identify where generative AI creates value for our customers and then come back to our AI experts who will step in to support. This is absolutely critical for us.”
At Stellantis, too, managers are a separate target, “because they are the ones who initiate change,” describes a group training manager. They benefit from the Generative AI for managers program, established with HEC. Other employees are improving their skills in prompting, or in certain use cases, professions and tools. Developers will soon learn to generate code. As for the experts, they are those in data, which companies have always been fighting over. Most of them train their AI and data teams due to the lack of skills at the end of the schools. Some of the 350 people in the AI Hub created in 2021 by Schneider Electric have specialized in generative AI. But there is still a lot of expertise among service providers: Groupe Rocher asks a Midjourney specialist to intervene when the communications department needs to generate images.
According to Valérie Perrier, senior account manager at LinkedIn, “In 2024, not taking the turn of generative AI has become a risk for an employee. HR has a crucial role to play in not letting a gap widen.” But training departments are struggling to keep up. Designing a course takes time, and here, everything is going too fast. So it is AI that sometimes creates them, searches the depths of the internet for new use cases, applies the fundamentals of pedagogy, produces exercises and their assessments. While many employees show an appetite for generative AI, others need to be encouraged to follow the training offered by their employer. The fear of not knowing how to use the tool or of having their job stolen is causing reluctance. “Acculturating as many people as possible reduces anxiety, analyse Siegfried de Preville. They are realizing that good use of AI cannot do without their critical thinking and business skills. Artificial intelligence can help you optimize your time, but it is essential to review and validate the results it provides. As with an intern. The question remains how to use these freed up working hours… Don’t worry, your employer is thinking about it.
The prompt engineer speaks into the ears of AIs
If all employees become prompters, does the “prompt engineer” that everyone is talking about exist? Do you really need to be a query expert to communicate with an AI? “Prompter is simple, there are lots of training courses on YouTube,” says Valérie Perrier, senior account manager at LinkedIn. In addition, generative AIs offer “prompt assistants” that write queries for you, and companies are equipping themselves with libraries of prompts, off-the-shelf queries ready to be used. Generating images or lines of code requires additional expertise compared to creating text. But this remains within the reach of non-initiates. On the other hand, “We will always need specialists to configure the model for complex tasks, says Flavien Parrel, Accenture’s industry of the future director. It is necessary to prepare the data that drives it and the learning parameters well, to formulate the instructions well and specify its expectations, to iterate to improve the responses. All this requires real expertise, prompt engineer is the job of the future, an extension of the job of data scientist. Knowing how to talk to machines may not be as easy as it seems.
1722455952
#Stellantis #Thales #Schneider #Electric.. #manufacturers #train #employees #generative
